New extension: access log based metrics

[kyessenov]

This is a proposal to add a new access logger that emits metrics using the Envoy metrics system. An example instance of this extension would be Istio standard metrics that are derived from the access log entries inside the proxy.

Note that the concern about the scalability does apply, because an access log value cardinality is unbounded. A few mitigation measures must be placed on this extension to be safe to use in production:

• the metrics scopes must have a limited lifetime;
• integration with the overload manager to drop the metrics with the excessive cardinality;

CC @jmarantz @briansonnenberg

CC #29681

[wbpcode]

But how we can manage the lifetime of these metrics? Seems these metrics may have a global or listener scope?

[kyessenov]

Yeah, they can be scoped to the listener or global. The lifetime for metrics needs to be decoupled from xDS parent resources because the cardinality depends on the traffic. That means the scope should be rotated periodically without the parent reload, for example.

[github-actions]

This issue has been automatically marked as stale because it has not had activity in the last 30 days. It will be closed in the next 7 days unless it is tagged "help wanted" or "no stalebot" or other activity occurs. Thank you for your contributions.

[github-actions]

This issue has been automatically closed because it has not had activity in the last 37 days. If this issue is still valid, please ping a maintainer and ask them to label it as "help wanted" or "no stalebot". Thank you for your contributions.

[github-actions]

This issue has been automatically marked as stale because it has not had activity in the last 30 days. It will be closed in the next 7 days unless it is tagged "help wanted" or "no stalebot" or other activity occurs. Thank you for your contributions.