fix(test): flaky `restart_sink_connector_continues_processing` integration test

[atharvalade]

Fix flaky restart_sink_connector_continues_processing integration test

Problem:

The test connectors::postgres::restart::restart_sink_connector_continues_processing is intermittently failing in CI:
FLAKY 2/4 [2.413s] integration::mod connectors::postgres::restart::restart_sink_connector_continues_processing

Root Cause:

fetch_rows_as in core/integration/tests/connectors/fixtures/postgres/sink.rs uses a sleep interval of DEFAULT_POLL_INTERVAL_MS / 5 (10ms), giving it a total polling budget of only ~1 second (100 attempts × 10ms). All other polling helpers (wait_for_sink_status, wait_for_table) use the full 50ms interval, giving them ~5 seconds.

After a connector restart, the sink needs to reconnect to Postgres and resume processing. On a loaded CI runner, 1 second is not always enough for the rows to appear, causing the test to fail with fewer rows than expected.

[atharvalade]

please assign this to me