Document istiod's logging scopes and how adjust their levels

[ndombroski]

The documentation is currently lacking any instruction on how to adjust the logging level of istiod's logs, which is a common need for users, especially because the data plane is incredibly verbose in larger Kubernetes clusters out-of-the-box.

Acceptance Criteria

• Each of istiod's logging scopes (default, delta, cds, etc.) are enumerated so users can reference which scopes exist.
• The docs instruct how to adjust an individual scope (e.g. --log_output_level=default:info,delta:warn)
• The docs instruct how to adjust all scopes (e.g. --log_output_level=all:warn)

[ndombroski]

If this sounds acceptable, I'd be happy to contribute this information, but I would need some guidance on where it would best fit. The existing telemetry section (https://istio.io/latest/docs/reference/config/telemetry/) seems to only pertain to proxy workloads, not istiod's telemetry.

[howardjohn]

Seems reasonable, but a few comments.

We should likely automate this, otherwise it will go out of sync very quickly. Or don't document the entire set of scopes, but instead how to find them (istioctl admin log)

because the data plane is incredibly verbose in larger Kubernetes clusters out-of-the-box.

None of this impacts access logs which is probably the 'data plane' part you are referring to? otherwise its the control plane?

[ndombroski]

Hey @howardjohn 👋 thanks for getting back to me.

That makes sense to me, documenting how to find the logs via istioctl instead sounds great.

And yes sorry I misspoke (mis-typed 😄), I had meant the control plane (istiod). Specifically for my org it was the delta scope that was way too high volume for our logging infra to support and I imagine other users will similarly want to tune the logging volume. So to clarify my last comment, the existing telemetry page seems to only cover the telemetry of proxy workloads (such as access logging) so I wasn't sure if inserting info about configuring istiod would fit there.

[kfaseela]

Hey @howardjohn 👋 thanks for getting back to me.

That makes sense to me, documenting how to find the logs via istioctl instead sounds great.

And yes sorry I misspoke (mis-typed 😄), I had meant the control plane (istiod). Specifically for my org it was the delta scope that was way too high volume for our logging infra to support and I imagine other users will similarly want to tune the logging volume. So to clarify my last comment, the existing telemetry page seems to only cover the telemetry of proxy workloads (such as access logging) so I wasn't sure if inserting info about configuring istiod would fit there.

@ndombroski feel free to contribute the PR as aligned with John's comments