You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The MeterProvider shutdown sequence is SdkMeterProvider ::Shutdown() -> opentelemetry_sdk::metrics::Pipeline::shutdown() -> PeriodicReader::shutdown() -> PushMetricsExporter::shutdown
So on shutdown,
the MetricReader will do one last collect-and-export cycle, and then stop, and set the exporter state to shutdown.
Any new Meters (and subsequently instruments from these meters) created after shutdown will be Noop.
However, the instruments created before the shutdown process can still record measurements. This applies as well to instruments that are created after the shutdown, if they are derived from meters that were established before the shutdown commenced. This behavior implies that these instruments are not aware of the shutdown state, continuing to consume and potentially increase memory with new measurements. This can be an issue with the long-running application.
Describe the solution you'd like:
The suggested approach would be
Stop recording all instruments after shutdown.
Cleanup the existing metrics for all the instruments.
Considered Alternatives
No response
Additional Context
No response
The text was updated successfully, but these errors were encountered:
Related Problems?
The MeterProvider shutdown sequence is
SdkMeterProvider ::Shutdown()
->opentelemetry_sdk::metrics::Pipeline::shutdown()
->PeriodicReader::shutdown()
->PushMetricsExporter::shutdown
So on shutdown,
However, the instruments created before the shutdown process can still record measurements. This applies as well to instruments that are created after the shutdown, if they are derived from meters that were established before the shutdown commenced. This behavior implies that these instruments are not aware of the shutdown state, continuing to consume and potentially increase memory with new measurements. This can be an issue with the long-running application.
Describe the solution you'd like:
The suggested approach would be
Considered Alternatives
No response
Additional Context
No response
The text was updated successfully, but these errors were encountered: