A data buffer for job metrics and logs

[sumerman]

Description

This PR introduces a Rust-based escape hatch for backend metrics (and, in the future, logs) data backend_telemetry_buffer.

Addresses #495

Merge requirements

Please take into account the following non-code changes that you may need to make with your PR:

• [ ] CHANGELOG entry for user-facing changes
• [ ] Updated the relevant documentation

[sumerman]

I validated the absence of serious memory leaks by running the following multiple times:

explain analyse SELECT _prom_ext.push_rec(ROW(now(), x)::_prom_ext.backend_telemetry_rec) FROM generate_series(1, 200000) x;
SELECT COUNT(*) FROM _prom_ext.pop_recs();

After executing the first line, the RAM consumption always jumped by about 15 MB and stayed, then reliably returned to its original reading at the end of the test.

[sumerman]

CALL run_job(:job_id); doesn't seem to be using the actual maintenance job infrastructure. In order to prove the buffer works inside an actual job worker I've repeated the integration test on a regular instance.

[sumerman]

Closing in favor of https://github.com/timescale/promscale_extension/pull/555