feat(crdb): experimental in-band query cancellation

[ecordell]

When a request context is canceled mid-query on the CRDB datastore, pgx's default behavior is to destroy the connection. Under load this drains the write pool and triggers the metastable death spiral described in #2576. A previous attempt to fix this using pgx's built-in CancelRequestContextWatcherHandler (pgwire cancel protocol) was reverted in #2434 because CRDB applies pgwire cancels asynchronously; a late-arriving cancel could kill the next query on the connection.

This PR implements in-band cancellation using CockroachDB's CANCEL QUERIES statement instead. A new pool.Canceler owns a small dedicated connection pool and a registry mapping each pooled write connection to its CRDB session_id (captured via SHOW session_id at connect). A custom ctxwatch.Handler fires on context cancellation and issues CANCEL QUERIES IF EXISTS (SELECT query_id FROM [SHOW CLUSTER STATEMENTS] WHERE session_id = '...') on a sibling connection, then blocks until that statement completes before releasing the original connection back to the pool. This sequencing eliminates the wrong-query race by construction: no cancellation can be in flight when the next query starts on the same session.

The feature is gated behind --datastore-experimental-crdb-query-cancellation (default off). When enabled, set --write-conn-acquisition-timeout=0; with connections no longer being destroyed on cancel, pool exhaustion under load is far less likely and the acquisition-timeout backpressure is no longer needed.

[codecov]

Codecov Report

❌ Patch coverage is 85.26316% with 28 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
internal/datastore/crdb/pool/canceler.go	74.69%	14 Missing and 6 partials ⚠️
internal/datastore/crdb/crdb.go	52.95%	6 Missing and 2 partials ⚠️

📢 Thoughts on this report? Let us know!