fix: resolve database timeout hotspots - #1385

Merged

ding113 merged 18 commits into

devfrom

fix/database-timeout-p0-p1

Aug 2, 2026

Commits on Aug 2, 2026

fix(session): align queries with identity indexes
ding113
committed
dc6e367
fix(replay): bound and coordinate expired payload cleanup
ding113
committed
5abca84
fix(dashboard): surface session and proxy status failures
ding113
committed
00fa1ae
feat(users): queue full statistics resets
ding113
committed
fe9b6b6
fix(migrations): roll out timeout indexes concurrently
ding113
committed
6e5f8ab
fix(proxy): keep durable replay winners visible on concurrent conflicts
ding113
committed
af545d8
feat(reset): preserve fixed 5h cost windows and recover ambiguous enqueues
ding113
committed
72be468
fix(sessions): normalize browser abort errors to stable cancellation code
ding113
committed
09a2feb
fix(ui): add exponential backoff and retry limit to reset polling
ding113
committed
0a19dfb
fix(reset-queue): guard failed-job handler and persist status safely
ding113
committed
8687d07
fix(reset-store): wait for Redis readiness before status operations
ding113
committed
21c3e96
fix(cost-cache): count scan failures in pipeline error results
ding113
committed
de0f56d
fix(api): use standard public detail for missing statistics reset
ding113
committed
9c91d45
fix(dashboard): harden statistics reset polling and dev queue safety
ding113
committed
a7019a6
fix(reset): delegate Redis readiness check to the reset queue
ding113
committed
406c83b
test(reset): align fixed 5h reset tests with queue-owned Redis readiness
ding113
committed
adf93c5
fix(replay): preserve live attachments on durable reuse
ding113
committed
f70d563
fix(message): prevent reserved session identities from aliasing physical ledger rows
ding113
committed
18c6e21

Navigation Menu

Uh oh!

fix: resolve database timeout hotspots - #1385

fix: resolve database timeout hotspots#1385
ding113 merged 18 commits into
devfrom
fix/database-timeout-p0-p1

Commits on Aug 2, 2026

fix(session): align queries with identity indexes

fix(replay): bound and coordinate expired payload cleanup

fix(dashboard): surface session and proxy status failures

feat(users): queue full statistics resets

fix(migrations): roll out timeout indexes concurrently

fix(proxy): keep durable replay winners visible on concurrent conflicts

feat(reset): preserve fixed 5h cost windows and recover ambiguous enqueues

fix(sessions): normalize browser abort errors to stable cancellation code

fix(ui): add exponential backoff and retry limit to reset polling

fix(reset-queue): guard failed-job handler and persist status safely

fix(reset-store): wait for Redis readiness before status operations

fix(cost-cache): count scan failures in pipeline error results

fix(api): use standard public detail for missing statistics reset

fix(dashboard): harden statistics reset polling and dev queue safety

fix(reset): delegate Redis readiness check to the reset queue

test(reset): align fixed 5h reset tests with queue-owned Redis readiness

fix(replay): preserve live attachments on durable reuse

fix(message): prevent reserved session identities from aliasing physical ledger rows