1
0
mirror of https://github.com/open-telemetry/opentelemetry-go.git synced 2025-01-22 03:38:42 +02:00
David Ashpole 6b251b804c
Allow configuring the exemplar filter on the metrics SDK (#5850)
Part of https://github.com/open-telemetry/opentelemetry-go/issues/5249

### Spec

https://opentelemetry.io/docs/specs/otel/metrics/sdk/#exemplarfilter

> The ExemplarFilter configuration MUST allow users to select between
one of the built-in ExemplarFilters. While ExemplarFilter determines
which measurements are eligible for becoming an Exemplar, the
ExemplarReservoir makes the final decision if a measurement becomes an
exemplar and is stored.

> The ExemplarFilter SHOULD be a configuration parameter of a
MeterProvider for an SDK. The default value SHOULD be TraceBased. The
filter configuration SHOULD follow the [environment variable
specification](https://opentelemetry.io/docs/specs/otel/configuration/sdk-environment-variables/#exemplar).

> An OpenTelemetry SDK MUST support the following filters:

> *
[AlwaysOn](https://opentelemetry.io/docs/specs/otel/metrics/sdk/#alwayson)
> *
[AlwaysOff](https://opentelemetry.io/docs/specs/otel/metrics/sdk/#alwaysoff)
> *
[TraceBased](https://opentelemetry.io/docs/specs/otel/metrics/sdk/#tracebased)

### Changes

* adds exemplar.AlwaysOffFilter, which is one of the required filters
from the SDK:
https://opentelemetry.io/docs/specs/otel/metrics/sdk/#alwaysoff
* adds `metric.WithExemplarFilter` as an option for the metrics SDK.
* moves handling of `OTEL_METRICS_EXEMPLAR_FILTER` to the same location
as config handling to make code easier to navigate.



dropReservoir can actually be removed, but I plan to do that in a
follow-up refactor, since it will be a large diff.

---------

Co-authored-by: Damien Mathieu <42@dmathieu.com>
Co-authored-by: Tyler Yahn <MrAlias@users.noreply.github.com>
Co-authored-by: Chester Cheung <cheung.zhy.csu@gmail.com>
2024-10-11 16:02:20 -04:00

144 lines
4.6 KiB
Go

// Copyright The OpenTelemetry Authors
// SPDX-License-Identifier: Apache-2.0
package metric // import "go.opentelemetry.io/otel/sdk/metric"
import (
"context"
"sync/atomic"
"go.opentelemetry.io/otel/internal/global"
"go.opentelemetry.io/otel/metric"
"go.opentelemetry.io/otel/metric/embedded"
"go.opentelemetry.io/otel/metric/noop"
"go.opentelemetry.io/otel/sdk/instrumentation"
)
// MeterProvider handles the creation and coordination of Meters. All Meters
// created by a MeterProvider will be associated with the same Resource, have
// the same Views applied to them, and have their produced metric telemetry
// passed to the configured Readers.
type MeterProvider struct {
embedded.MeterProvider
pipes pipelines
meters cache[instrumentation.Scope, *meter]
forceFlush, shutdown func(context.Context) error
stopped atomic.Bool
}
// Compile-time check MeterProvider implements metric.MeterProvider.
var _ metric.MeterProvider = (*MeterProvider)(nil)
// NewMeterProvider returns a new and configured MeterProvider.
//
// By default, the returned MeterProvider is configured with the default
// Resource and no Readers. Readers cannot be added after a MeterProvider is
// created. This means the returned MeterProvider, one created with no
// Readers, will perform no operations.
func NewMeterProvider(options ...Option) *MeterProvider {
conf := newConfig(options)
flush, sdown := conf.readerSignals()
mp := &MeterProvider{
pipes: newPipelines(conf.res, conf.readers, conf.views, conf.exemplarFilter),
forceFlush: flush,
shutdown: sdown,
}
// Log after creation so all readers show correctly they are registered.
global.Info("MeterProvider created",
"Resource", conf.res,
"Readers", conf.readers,
"Views", len(conf.views),
)
return mp
}
// Meter returns a Meter with the given name and configured with options.
//
// The name should be the name of the instrumentation scope creating
// telemetry. This name may be the same as the instrumented code only if that
// code provides built-in instrumentation.
//
// Calls to the Meter method after Shutdown has been called will return Meters
// that perform no operations.
//
// This method is safe to call concurrently.
func (mp *MeterProvider) Meter(name string, options ...metric.MeterOption) metric.Meter {
if name == "" {
global.Warn("Invalid Meter name.", "name", name)
}
if mp.stopped.Load() {
return noop.Meter{}
}
c := metric.NewMeterConfig(options...)
s := instrumentation.Scope{
Name: name,
Version: c.InstrumentationVersion(),
SchemaURL: c.SchemaURL(),
}
global.Info("Meter created",
"Name", s.Name,
"Version", s.Version,
"SchemaURL", s.SchemaURL,
)
return mp.meters.Lookup(s, func() *meter {
return newMeter(s, mp.pipes)
})
}
// ForceFlush flushes all pending telemetry.
//
// This method honors the deadline or cancellation of ctx. An appropriate
// error will be returned in these situations. There is no guaranteed that all
// telemetry be flushed or all resources have been released in these
// situations.
//
// ForceFlush calls ForceFlush(context.Context) error
// on all Readers that implements this method.
//
// This method is safe to call concurrently.
func (mp *MeterProvider) ForceFlush(ctx context.Context) error {
if mp.forceFlush != nil {
return mp.forceFlush(ctx)
}
return nil
}
// Shutdown shuts down the MeterProvider flushing all pending telemetry and
// releasing any held computational resources.
//
// This call is idempotent. The first call will perform all flush and
// releasing operations. Subsequent calls will perform no action and will
// return an error stating this.
//
// Measurements made by instruments from meters this MeterProvider created
// will not be exported after Shutdown is called.
//
// This method honors the deadline or cancellation of ctx. An appropriate
// error will be returned in these situations. There is no guaranteed that all
// telemetry be flushed or all resources have been released in these
// situations.
//
// This method is safe to call concurrently.
func (mp *MeterProvider) Shutdown(ctx context.Context) error {
// Even though it may seem like there is a synchronization issue between the
// call to `Store` and checking `shutdown`, the Go concurrency model ensures
// that is not the case, as all the atomic operations executed in a program
// behave as though executed in some sequentially consistent order. This
// definition provides the same semantics as C++'s sequentially consistent
// atomics and Java's volatile variables.
// See https://go.dev/ref/mem#atomic and https://pkg.go.dev/sync/atomic.
mp.stopped.Store(true)
if mp.shutdown != nil {
return mp.shutdown(ctx)
}
return nil
}