Reduce performance logging to DEBUG

[michaelkaye]

Pull Request Checklist

• Pull request is based on the develop branch
• Pull request includes a changelog file. The entry should:
  • Be a short description of your change which makes sense to users. "Fixed a bug that prevented receiving messages from other servers." instead of "Moved X method from EventStore to EventWorkerStore.".
  • Use markdown where necessary, mostly for code blocks.
  • End with either a period (.) or an exclamation mark (!).
  • Start with a capital letter.
• Pull request includes a sign off
• Code style is correct (run the linters)

This PR reduces the default logging level of verbose performance monitoring counters from INFO to DEBUG.

Leo even wrote along side this code:

        # TODO(paul): These can eventually be removed once the metrics code
        #   is running in mainline, and we have some nice monitoring frontends
        #   to watch it

So there's definitely a case to rip this out entirely - reducing to DEBUG seems safe.

These metrics are available more generally in the prometheus metrics (eg synapse_background_process_db_txn_duration_seconds) which when looking at performance changes over time is probably a better way to access this data.

This change originated in feedback that these messages are confusing to server administrators, especially at the INFO level.

[richvdh]

lgtm, though I wonder if we should get rid of all the spurious _txn_perf_counters stuff.

[michaelkaye]

lgtm, though I wonder if we should get rid of all the spurious _txn_perf_counters stuff.

I would not be opposed to that, but maybe a larger change. I'll open an issue

[richvdh]

quick reminder that changelogs should be in the imperative mood: Reduce log level to DEBUG for synapse.storage.TIME.