1
0
mirror of https://github.com/open-telemetry/opentelemetry-go.git synced 2024-12-14 10:13:10 +02:00
opentelemetry-go/exporters/otlp/grpcconnection.go

272 lines
7.1 KiB
Go
Raw Normal View History

// Copyright The OpenTelemetry Authors
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package otlp // import "go.opentelemetry.io/otel/exporters/otlp"
import (
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
"context"
"math/rand"
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
"sync"
"sync/atomic"
"time"
"unsafe"
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
"google.golang.org/grpc"
"google.golang.org/grpc/metadata"
)
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
type grpcConnection struct {
// Ensure pointer is 64-bit aligned for atomic operations on both 32 and 64 bit machines.
lastConnectErrPtr unsafe.Pointer
// mu protects the connection as it is accessed by the
// exporter goroutines and background connection goroutine
mu sync.Mutex
cc *grpc.ClientConn
// these fields are read-only after constructor is finished
Split connection management away from exporter (#1369) * Split protocol handling away from exporter This commits adds a ProtocolDriver interface, which the exporter will use to connect to the collector and send both metrics and traces to it. That way, the Exporter type is free from dealing with any connection/protocol details, as this business is taken over by the implementations of the ProtocolDriver interface. The gRPC code from the exporter is moved into the implementation of ProtocolDriver. Currently it only maintains a single connection, just as the Exporter used to do. With the split, most of the Exporter options became actually gRPC connection manager's options. Currently the only option that remained to be Exporter's is about setting the export kind selector. * Update changelog * Increase the test coverage of GRPC driver * Do not close a channel with multiple senders The disconnected channel can be used for sending by multiple goroutines (for example, by metric controller and span processor), so this channel should not be closed at all. Dropping this line closes a race between closing a channel and sending to it. * Simplify new connection handler The callbacks never return an error, so drop the return type from it. * Access clients under a lock The client may change as a result on reconnection in background, so guard against a racy access. * Simplify the GRPC driver a bit The config type was exported earlier to have a consistent way of configuring the driver, when also the multiple connection driver would appear. Since we are not going to add a multiple connection driver, pass the options directly to the driver constructor. Also shorten the name of the constructor to `NewGRPCDriver`. * Merge common gRPC code back into the driver The common code was supposed to be shared between single connection driver and multiple connection driver, but since the latter won't be happening, it makes no sense to keep the not-so-common code in a separate file. Also drop some abstraction too. * Rename the file with gRPC driver implementation * Update changelog * Sleep for a second to trigger the timeout Sometimes CI has it's better moments, so it's blazing fast and manages to finish shutting the exporter down within the 1 microsecond timeout. * Increase the timeout for shutting down the exporter One millisecond is quite short, and I was getting failures locally or in CI: go test ./... + race in ./exporters/otlp 2020/12/14 18:27:54 rpc error: code = Canceled desc = context canceled 2020/12/14 18:27:54 context deadline exceeded --- FAIL: TestNewExporter_withMultipleAttributeTypes (0.37s) otlp_integration_test.go:541: resource span count: got 0, want 1 FAIL FAIL go.opentelemetry.io/otel/exporters/otlp 5.278s or go test ./... + coverage in ./exporters/otlp 2020/12/14 17:41:16 rpc error: code = Canceled desc = context canceled 2020/12/14 17:41:16 exporter disconnected --- FAIL: TestNewExporter_endToEnd (1.53s) --- FAIL: TestNewExporter_endToEnd/WithCompressor (0.41s) otlp_integration_test.go:246: span counts: got 3, want 4 2020/12/14 17:41:18 context canceled FAIL coverage: 35.3% of statements in ./... FAIL go.opentelemetry.io/otel/exporters/otlp 4.753s * Shut down the providers in end to end test This is to make sure that all batched spans are actually flushed before closing the exporter.
2020-12-21 22:49:45 +02:00
c grpcConnectionConfig
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
metadata metadata.MD
Split connection management away from exporter (#1369) * Split protocol handling away from exporter This commits adds a ProtocolDriver interface, which the exporter will use to connect to the collector and send both metrics and traces to it. That way, the Exporter type is free from dealing with any connection/protocol details, as this business is taken over by the implementations of the ProtocolDriver interface. The gRPC code from the exporter is moved into the implementation of ProtocolDriver. Currently it only maintains a single connection, just as the Exporter used to do. With the split, most of the Exporter options became actually gRPC connection manager's options. Currently the only option that remained to be Exporter's is about setting the export kind selector. * Update changelog * Increase the test coverage of GRPC driver * Do not close a channel with multiple senders The disconnected channel can be used for sending by multiple goroutines (for example, by metric controller and span processor), so this channel should not be closed at all. Dropping this line closes a race between closing a channel and sending to it. * Simplify new connection handler The callbacks never return an error, so drop the return type from it. * Access clients under a lock The client may change as a result on reconnection in background, so guard against a racy access. * Simplify the GRPC driver a bit The config type was exported earlier to have a consistent way of configuring the driver, when also the multiple connection driver would appear. Since we are not going to add a multiple connection driver, pass the options directly to the driver constructor. Also shorten the name of the constructor to `NewGRPCDriver`. * Merge common gRPC code back into the driver The common code was supposed to be shared between single connection driver and multiple connection driver, but since the latter won't be happening, it makes no sense to keep the not-so-common code in a separate file. Also drop some abstraction too. * Rename the file with gRPC driver implementation * Update changelog * Sleep for a second to trigger the timeout Sometimes CI has it's better moments, so it's blazing fast and manages to finish shutting the exporter down within the 1 microsecond timeout. * Increase the timeout for shutting down the exporter One millisecond is quite short, and I was getting failures locally or in CI: go test ./... + race in ./exporters/otlp 2020/12/14 18:27:54 rpc error: code = Canceled desc = context canceled 2020/12/14 18:27:54 context deadline exceeded --- FAIL: TestNewExporter_withMultipleAttributeTypes (0.37s) otlp_integration_test.go:541: resource span count: got 0, want 1 FAIL FAIL go.opentelemetry.io/otel/exporters/otlp 5.278s or go test ./... + coverage in ./exporters/otlp 2020/12/14 17:41:16 rpc error: code = Canceled desc = context canceled 2020/12/14 17:41:16 exporter disconnected --- FAIL: TestNewExporter_endToEnd (1.53s) --- FAIL: TestNewExporter_endToEnd/WithCompressor (0.41s) otlp_integration_test.go:246: span counts: got 3, want 4 2020/12/14 17:41:18 context canceled FAIL coverage: 35.3% of statements in ./... FAIL go.opentelemetry.io/otel/exporters/otlp 4.753s * Shut down the providers in end to end test This is to make sure that all batched spans are actually flushed before closing the exporter.
2020-12-21 22:49:45 +02:00
newConnectionHandler func(cc *grpc.ClientConn)
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
// these channels are created once
disconnectedCh chan bool
backgroundConnectionDoneCh chan struct{}
stopCh chan struct{}
// this is for tests, so they can replace the closing
// routine without a worry of modifying some global variable
// or changing it back to original after the test is done
closeBackgroundConnectionDoneCh func(ch chan struct{})
}
Split connection management away from exporter (#1369) * Split protocol handling away from exporter This commits adds a ProtocolDriver interface, which the exporter will use to connect to the collector and send both metrics and traces to it. That way, the Exporter type is free from dealing with any connection/protocol details, as this business is taken over by the implementations of the ProtocolDriver interface. The gRPC code from the exporter is moved into the implementation of ProtocolDriver. Currently it only maintains a single connection, just as the Exporter used to do. With the split, most of the Exporter options became actually gRPC connection manager's options. Currently the only option that remained to be Exporter's is about setting the export kind selector. * Update changelog * Increase the test coverage of GRPC driver * Do not close a channel with multiple senders The disconnected channel can be used for sending by multiple goroutines (for example, by metric controller and span processor), so this channel should not be closed at all. Dropping this line closes a race between closing a channel and sending to it. * Simplify new connection handler The callbacks never return an error, so drop the return type from it. * Access clients under a lock The client may change as a result on reconnection in background, so guard against a racy access. * Simplify the GRPC driver a bit The config type was exported earlier to have a consistent way of configuring the driver, when also the multiple connection driver would appear. Since we are not going to add a multiple connection driver, pass the options directly to the driver constructor. Also shorten the name of the constructor to `NewGRPCDriver`. * Merge common gRPC code back into the driver The common code was supposed to be shared between single connection driver and multiple connection driver, but since the latter won't be happening, it makes no sense to keep the not-so-common code in a separate file. Also drop some abstraction too. * Rename the file with gRPC driver implementation * Update changelog * Sleep for a second to trigger the timeout Sometimes CI has it's better moments, so it's blazing fast and manages to finish shutting the exporter down within the 1 microsecond timeout. * Increase the timeout for shutting down the exporter One millisecond is quite short, and I was getting failures locally or in CI: go test ./... + race in ./exporters/otlp 2020/12/14 18:27:54 rpc error: code = Canceled desc = context canceled 2020/12/14 18:27:54 context deadline exceeded --- FAIL: TestNewExporter_withMultipleAttributeTypes (0.37s) otlp_integration_test.go:541: resource span count: got 0, want 1 FAIL FAIL go.opentelemetry.io/otel/exporters/otlp 5.278s or go test ./... + coverage in ./exporters/otlp 2020/12/14 17:41:16 rpc error: code = Canceled desc = context canceled 2020/12/14 17:41:16 exporter disconnected --- FAIL: TestNewExporter_endToEnd (1.53s) --- FAIL: TestNewExporter_endToEnd/WithCompressor (0.41s) otlp_integration_test.go:246: span counts: got 3, want 4 2020/12/14 17:41:18 context canceled FAIL coverage: 35.3% of statements in ./... FAIL go.opentelemetry.io/otel/exporters/otlp 4.753s * Shut down the providers in end to end test This is to make sure that all batched spans are actually flushed before closing the exporter.
2020-12-21 22:49:45 +02:00
func newGRPCConnection(c grpcConnectionConfig, handler func(cc *grpc.ClientConn)) *grpcConnection {
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
conn := new(grpcConnection)
conn.newConnectionHandler = handler
conn.c = c
if len(conn.c.headers) > 0 {
conn.metadata = metadata.New(conn.c.headers)
}
conn.closeBackgroundConnectionDoneCh = func(ch chan struct{}) {
close(ch)
}
return conn
}
func (oc *grpcConnection) startConnection(ctx context.Context) {
oc.stopCh = make(chan struct{})
oc.disconnectedCh = make(chan bool)
oc.backgroundConnectionDoneCh = make(chan struct{})
if err := oc.connect(ctx); err == nil {
oc.setStateConnected()
} else {
oc.setStateDisconnected(err)
}
go oc.indefiniteBackgroundConnection()
}
func (oc *grpcConnection) lastConnectError() error {
errPtr := (*error)(atomic.LoadPointer(&oc.lastConnectErrPtr))
if errPtr == nil {
return nil
}
return *errPtr
}
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
func (oc *grpcConnection) saveLastConnectError(err error) {
var errPtr *error
if err != nil {
errPtr = &err
}
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
atomic.StorePointer(&oc.lastConnectErrPtr, unsafe.Pointer(errPtr))
}
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
func (oc *grpcConnection) setStateDisconnected(err error) {
oc.saveLastConnectError(err)
select {
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
case oc.disconnectedCh <- true:
default:
}
Split connection management away from exporter (#1369) * Split protocol handling away from exporter This commits adds a ProtocolDriver interface, which the exporter will use to connect to the collector and send both metrics and traces to it. That way, the Exporter type is free from dealing with any connection/protocol details, as this business is taken over by the implementations of the ProtocolDriver interface. The gRPC code from the exporter is moved into the implementation of ProtocolDriver. Currently it only maintains a single connection, just as the Exporter used to do. With the split, most of the Exporter options became actually gRPC connection manager's options. Currently the only option that remained to be Exporter's is about setting the export kind selector. * Update changelog * Increase the test coverage of GRPC driver * Do not close a channel with multiple senders The disconnected channel can be used for sending by multiple goroutines (for example, by metric controller and span processor), so this channel should not be closed at all. Dropping this line closes a race between closing a channel and sending to it. * Simplify new connection handler The callbacks never return an error, so drop the return type from it. * Access clients under a lock The client may change as a result on reconnection in background, so guard against a racy access. * Simplify the GRPC driver a bit The config type was exported earlier to have a consistent way of configuring the driver, when also the multiple connection driver would appear. Since we are not going to add a multiple connection driver, pass the options directly to the driver constructor. Also shorten the name of the constructor to `NewGRPCDriver`. * Merge common gRPC code back into the driver The common code was supposed to be shared between single connection driver and multiple connection driver, but since the latter won't be happening, it makes no sense to keep the not-so-common code in a separate file. Also drop some abstraction too. * Rename the file with gRPC driver implementation * Update changelog * Sleep for a second to trigger the timeout Sometimes CI has it's better moments, so it's blazing fast and manages to finish shutting the exporter down within the 1 microsecond timeout. * Increase the timeout for shutting down the exporter One millisecond is quite short, and I was getting failures locally or in CI: go test ./... + race in ./exporters/otlp 2020/12/14 18:27:54 rpc error: code = Canceled desc = context canceled 2020/12/14 18:27:54 context deadline exceeded --- FAIL: TestNewExporter_withMultipleAttributeTypes (0.37s) otlp_integration_test.go:541: resource span count: got 0, want 1 FAIL FAIL go.opentelemetry.io/otel/exporters/otlp 5.278s or go test ./... + coverage in ./exporters/otlp 2020/12/14 17:41:16 rpc error: code = Canceled desc = context canceled 2020/12/14 17:41:16 exporter disconnected --- FAIL: TestNewExporter_endToEnd (1.53s) --- FAIL: TestNewExporter_endToEnd/WithCompressor (0.41s) otlp_integration_test.go:246: span counts: got 3, want 4 2020/12/14 17:41:18 context canceled FAIL coverage: 35.3% of statements in ./... FAIL go.opentelemetry.io/otel/exporters/otlp 4.753s * Shut down the providers in end to end test This is to make sure that all batched spans are actually flushed before closing the exporter.
2020-12-21 22:49:45 +02:00
oc.newConnectionHandler(nil)
}
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
func (oc *grpcConnection) setStateConnected() {
oc.saveLastConnectError(nil)
}
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
func (oc *grpcConnection) connected() bool {
return oc.lastConnectError() == nil
}
const defaultConnReattemptPeriod = 10 * time.Second
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
func (oc *grpcConnection) indefiniteBackgroundConnection() {
defer func() {
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
oc.closeBackgroundConnectionDoneCh(oc.backgroundConnectionDoneCh)
}()
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
connReattemptPeriod := oc.c.reconnectionPeriod
if connReattemptPeriod <= 0 {
connReattemptPeriod = defaultConnReattemptPeriod
}
// No strong seeding required, nano time can
// already help with pseudo uniqueness.
rng := rand.New(rand.NewSource(time.Now().UnixNano() + rand.Int63n(1024)))
// maxJitterNanos: 70% of the connectionReattemptPeriod
maxJitterNanos := int64(0.7 * float64(connReattemptPeriod))
for {
// Otherwise these will be the normal scenarios to enable
// reconnection if we trip out.
// 1. If we've stopped, return entirely
// 2. Otherwise block until we are disconnected, and
// then retry connecting
select {
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
case <-oc.stopCh:
return
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
case <-oc.disconnectedCh:
// Quickly check if we haven't stopped at the
// same time.
select {
case <-oc.stopCh:
return
default:
}
// Normal scenario that we'll wait for
}
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
if err := oc.connect(context.Background()); err == nil {
oc.setStateConnected()
} else {
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
oc.setStateDisconnected(err)
}
// Apply some jitter to avoid lockstep retrials of other
// collector-exporters. Lockstep retrials could result in an
// innocent DDOS, by clogging the machine's resources and network.
jitter := time.Duration(rng.Int63n(maxJitterNanos))
select {
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
case <-oc.stopCh:
return
case <-time.After(connReattemptPeriod + jitter):
}
}
}
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
func (oc *grpcConnection) connect(ctx context.Context) error {
cc, err := oc.dialToCollector(ctx)
if err != nil {
return err
}
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
oc.setConnection(cc)
Split connection management away from exporter (#1369) * Split protocol handling away from exporter This commits adds a ProtocolDriver interface, which the exporter will use to connect to the collector and send both metrics and traces to it. That way, the Exporter type is free from dealing with any connection/protocol details, as this business is taken over by the implementations of the ProtocolDriver interface. The gRPC code from the exporter is moved into the implementation of ProtocolDriver. Currently it only maintains a single connection, just as the Exporter used to do. With the split, most of the Exporter options became actually gRPC connection manager's options. Currently the only option that remained to be Exporter's is about setting the export kind selector. * Update changelog * Increase the test coverage of GRPC driver * Do not close a channel with multiple senders The disconnected channel can be used for sending by multiple goroutines (for example, by metric controller and span processor), so this channel should not be closed at all. Dropping this line closes a race between closing a channel and sending to it. * Simplify new connection handler The callbacks never return an error, so drop the return type from it. * Access clients under a lock The client may change as a result on reconnection in background, so guard against a racy access. * Simplify the GRPC driver a bit The config type was exported earlier to have a consistent way of configuring the driver, when also the multiple connection driver would appear. Since we are not going to add a multiple connection driver, pass the options directly to the driver constructor. Also shorten the name of the constructor to `NewGRPCDriver`. * Merge common gRPC code back into the driver The common code was supposed to be shared between single connection driver and multiple connection driver, but since the latter won't be happening, it makes no sense to keep the not-so-common code in a separate file. Also drop some abstraction too. * Rename the file with gRPC driver implementation * Update changelog * Sleep for a second to trigger the timeout Sometimes CI has it's better moments, so it's blazing fast and manages to finish shutting the exporter down within the 1 microsecond timeout. * Increase the timeout for shutting down the exporter One millisecond is quite short, and I was getting failures locally or in CI: go test ./... + race in ./exporters/otlp 2020/12/14 18:27:54 rpc error: code = Canceled desc = context canceled 2020/12/14 18:27:54 context deadline exceeded --- FAIL: TestNewExporter_withMultipleAttributeTypes (0.37s) otlp_integration_test.go:541: resource span count: got 0, want 1 FAIL FAIL go.opentelemetry.io/otel/exporters/otlp 5.278s or go test ./... + coverage in ./exporters/otlp 2020/12/14 17:41:16 rpc error: code = Canceled desc = context canceled 2020/12/14 17:41:16 exporter disconnected --- FAIL: TestNewExporter_endToEnd (1.53s) --- FAIL: TestNewExporter_endToEnd/WithCompressor (0.41s) otlp_integration_test.go:246: span counts: got 3, want 4 2020/12/14 17:41:18 context canceled FAIL coverage: 35.3% of statements in ./... FAIL go.opentelemetry.io/otel/exporters/otlp 4.753s * Shut down the providers in end to end test This is to make sure that all batched spans are actually flushed before closing the exporter.
2020-12-21 22:49:45 +02:00
oc.newConnectionHandler(cc)
return nil
Another batch of cleanups in otlp exporter (#1357) * Move connection logic into grpcConnection object If we will need to maintain more than one connection in future, this splitting off will come in handy. Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com> * Make another channel a signal channel There is another channel that serves as a one-time signal, where channel's data type does not matter. * Reorder and document connection members This is to make clear that the lock is guarding only the connection since it can be changed by multiple goroutines, and other members are either atomic or read-only. * Move stop signal into connection The stop channel was rather useless on the exporter side - the primary reason for existence of this channel is to stop a background reconnecting goroutine. Since the goroutine lives entirely within grpcConnection object, move the stop channel here. Also expose a function to unify the stop channel with the context cancellation, so exporter can use it without knowing anything about stop channels. Also make export functions a bit more consistent. * Do not run reconnection routine when being stopped too It's possible that both disconnected channel and stop channel will be triggered around the same time, so the goroutine is as likely to start reconnecting as to return from the goroutine. Make sure we return if the stop channel is closed. * Nil clients on connection error Set clients to nil on connection error, so we don't try to send the data over a bad connection, but return a "no client" error immediately. * Do not call new connection handler within critical section It's rather risky to call a callback coming from outside within a critical section. Move it out. * Add context parameter to connection routines Connecting to the collector may also take its time, so it can be useful in some cases to pass a context with a deadline. Currently we just pass a background context, so this commit does not really change any behavior. The follow-up commits will make a use of it, though. * Add context parameter to NewExporter and Start It makes it possible to limit the time spent on connecting to the collector. * Stop connecting on shutdown Dialling to grpc service ignored the closing of the stop channel, but this can be easily changed. * Close connection after background is shut down That way we can make sure that there won't be a window between closing a connection and waiting for the background goroutine to return, where the new connection could be established. * Remove unnecessary nil check This member is never nil, unless the Exporter is created like &Exporter{}, which is not a thing we support anyway. * Update changelog Co-authored-by: Stefan Prisca <stefan.prisca@gmail.com>
2020-11-24 21:50:05 +02:00
}
// setConnection sets cc as the client connection and returns true if
// the connection state changed.
func (oc *grpcConnection) setConnection(cc *grpc.ClientConn) bool {
oc.mu.Lock()
defer oc.mu.Unlock()
// If previous clientConn is same as the current then just return.
// This doesn't happen right now as this func is only called with new ClientConn.
// It is more about future-proofing.
if oc.cc == cc {
return false
}
// If the previous clientConn was non-nil, close it
if oc.cc != nil {
_ = oc.cc.Close()
}
oc.cc = cc
return true
}
func (oc *grpcConnection) dialToCollector(ctx context.Context) (*grpc.ClientConn, error) {
addr := oc.c.collectorAddr
dialOpts := []grpc.DialOption{}
if oc.c.grpcServiceConfig != "" {
dialOpts = append(dialOpts, grpc.WithDefaultServiceConfig(oc.c.grpcServiceConfig))
}
if oc.c.clientCredentials != nil {
dialOpts = append(dialOpts, grpc.WithTransportCredentials(oc.c.clientCredentials))
} else if oc.c.canDialInsecure {
dialOpts = append(dialOpts, grpc.WithInsecure())
}
if oc.c.compressor != "" {
dialOpts = append(dialOpts, grpc.WithDefaultCallOptions(grpc.UseCompressor(oc.c.compressor)))
}
if len(oc.c.grpcDialOptions) != 0 {
dialOpts = append(dialOpts, oc.c.grpcDialOptions...)
}
ctx, cancel := oc.contextWithStop(ctx)
defer cancel()
ctx = oc.contextWithMetadata(ctx)
return grpc.DialContext(ctx, addr, dialOpts...)
}
func (oc *grpcConnection) contextWithMetadata(ctx context.Context) context.Context {
if oc.metadata.Len() > 0 {
return metadata.NewOutgoingContext(ctx, oc.metadata)
}
return ctx
}
func (oc *grpcConnection) shutdown(ctx context.Context) error {
close(oc.stopCh)
// Ensure that the backgroundConnector returns
select {
case <-oc.backgroundConnectionDoneCh:
case <-ctx.Done():
return ctx.Err()
}
oc.mu.Lock()
cc := oc.cc
oc.cc = nil
oc.mu.Unlock()
if cc != nil {
return cc.Close()
}
return nil
}
func (oc *grpcConnection) contextWithStop(ctx context.Context) (context.Context, context.CancelFunc) {
// Unify the parent context Done signal with the connection's
// stop channel.
ctx, cancel := context.WithCancel(ctx)
go func(ctx context.Context, cancel context.CancelFunc) {
select {
case <-ctx.Done():
// Nothing to do, either cancelled or deadline
// happened.
case <-oc.stopCh:
cancel()
}
}(ctx, cancel)
return ctx, cancel
}