use standard Envoy interval reporting for telemetry #43763
Comments
|
@kyessenov, I would like to contribute here, could you please provide me with some more details if you have or any entry point in Istio code from where I can start? |
|
buffer_flush_interval ? this parameter should probably be placed under the Provider associated with the log |
|
@kyessenov do you means |
|
@zirain Yeah, and there is another one for HTTP after envoyproxy/envoy#25824. I think it's better than having our stats filters doing the same thing. |
|
are you talking about promte tcp_reporting_duration to upstream and resue it? |
|
We don't need to implement upstream - it's done by others. We should align our APIs to upstream better - so that the reporting duration is consistent across all sinks. E.g. the same reporting duration knos applies for OTel, stats, etc. |
|
make senses |
istio-policy-bot
added
the
lifecycle/stale
istio-policy-bot
added
the
lifecycle/automatically-closed
istio-policy-bot
removed
the
lifecycle/stale
|
This is future looking, for Otel. |
istio-policy-bot
added
the
lifecycle/stale
|
🚧 This issue or pull request has been closed due to not having had activity from an Istio team member since 2023-06-21. If you feel this issue or pull request deserves attention, please reopen the issue. Please see this wiki page for more information. Thank you for your contributions. Created by the issue and PR lifecycle manager. |
Envoy has added a standard
access_log_flush_intervalfeature to access logs. We should remove custom Istio code duplicating this feature and switch to the upstream way - this will help eventual transition to the standard OTLP and telemetry sinks.A possible issue is that TCP stats sets it to 15s, but the envoy option applies to all sinks at the same time.
The text was updated successfully, but these errors were encountered: