fix(metrics, logs): Ensure TelemetryBuffer uses own thread for both size based and periodic flushing

[sl0thentr0py]

Before this change, the add_item triggered size based flush on the caller thread which is absolutely incorrect behavior.

Changes to ThreadedPeriodicWorker:

• a dedicated @thread_mutex for syncing its own internal state
• a condition variable @wake_condition that is used for wakeing the thread from outside (like size based flushing) - otherwise defaults to the interval which makes it backwards compatible
• another condition variable @idle_condition used for the final flush that needs to wait for any running task to finish
• two new methods wake and wait_for_idle

These are then used in TelemetryEventBuffer:

• run and flush are now two different methods, both run on the buffer thread
• flush waits for a max 2 seconds timeout otherwise returns because we don't want to block too long

Issues

• Resolves: Telemetry buffer improvements #3091
• Resolves: RUBY-214

[sl0thentr0py]

reviewing this with whitespace off recommended

[cursor]

Stale Bugbot comment from a previous run.

[sl0thentr0py]

moved this here to avoid logger patch leaking into the other test

[cursor]

Stale Bugbot comment from a previous run.

[solnic]

I honestly think that at this point we'd be better off using concurrent-ruby's primitives for this stuff, we already depend on that gem anyway.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit f875712. Configure here.}

[sl0thentr0py]

ran sol pro on this for deadlock analysis
https://gist.github.com/sl0thentr0py/5be6b2ff7adee8508d86f5315d3e142e

[linear-code]

RUBY-214