feat(reliability): configurable DB pool tuning + pre_ping/recycle#929
Open
seonghobae wants to merge 1 commit into
Open
feat(reliability): configurable DB pool tuning + pre_ping/recycle#929seonghobae wants to merge 1 commit into
seonghobae wants to merge 1 commit into
Conversation
Both async engines were created with no pool configuration: default sizing under enterprise concurrency is a silent bottleneck, stale pooled connections surface as user-facing errors after a DB restart/failover, and there was no operator knob to tune any of it. - db/session.py: shared _engine_kwargs() for the primary and read-only engines. pool_pre_ping (default on) validates connections at checkout; pool_recycle (default 1800s) replaces aged connections before server-side idle timeouts kill them. - core/config.py: DB_POOL_SIZE / DB_MAX_OVERFLOW / DB_POOL_TIMEOUT_SECONDS default to None and are only passed when set, so unset config keeps exact SQLAlchemy defaults and stays compatible with SQLite pools in tests. - tests/test_db_session.py: extended with default (pre_ping on, no sizing passed) and configured-sizing assertions; existing replica-URL tests kept. Verified locally: test_db_session/test_db 4 passed; search/emails/bootstrap regressions 87 passed. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RkKdtHRLG4wSLh6PVsp8J
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Gap
Both async engines (
db/session.py) were created with no pool configuration (reliability audit #9): default sizing is a silent bottleneck under enterprise concurrency, a restarted/failed-over Postgres surfaces stale pooled connections as user-facing errors, and operators had no knob.What
_engine_kwargs()shared by primary + read-only engines:pool_pre_ping(default on) — validates connections at checkout; a DB restart/failover no longer causes stale-connection 500s.pool_recycle(default 1800s) — replaces aged connections before server-side idle timeouts kill them.DB_POOL_SIZE/DB_MAX_OVERFLOW/DB_POOL_TIMEOUT_SECONDS— defaultNone→ not passed → exact SQLAlchemy defaults (zero sizing change until an operator sets them; also keeps SQLite test pools compatible).Verification (local — CI frozen org-wide)
tests/test_db_session.py: default behavior (pre_ping on, recycle 1800, no sizing kwargs passed) + configured sizing (20/10/5/900/pre_ping off) + existing replica-URL tests updated for the new kwargs → 4 passed.test_search+test_emails_api+test_bootstrap_db→ 87 passed.Notes
This pairs with the HA drill docs (
postgresql-physical-replication.md): pre_ping is what makes a replica promotion recoverable without a backend restart.🤖 Generated with Claude Code