perf(ci): speed up unit tests with sharding + container scoping by nicktrn · Pull Request #3855 · triggerdotdev/trigger.dev

test(testcontainers): add local timing harness, docs, and TESTCONTAINERS_TIMING gate

nicktrn

committed

perf(testcontainers): force-kill teardown + worker-scope clickhouse/redis/postgres containers

nicktrn

committed

test(run-engine): make ttl consumer waits event-driven instead of fixed 13s sleeps

nicktrn

committed

perf(testcontainers): worker-scope containerTest (compose pg/redis/clickhouse isolation)

nicktrn

committed

perf(testcontainers): scope minio + tails; carve out replicationContainerTest for logical replication

nicktrn

committed

test(webapp): move logical-replication suites to replicationContainerTest

nicktrn

committed

test(run-engine): make waitpoints waits event-driven (fix the original flaky race)

nicktrn

committed

test(testcontainers): add opt-in CI-resource-limit mode for local flake repro

nicktrn

committed

perf(testcontainers): replicationContainerTest only needs per-test postgres (share redis+clickhouse)

nicktrn

committed

fix(testcontainers): per-test redis for background-worker tests + faster redis boot

nicktrn

committed

perf(ci): duration-weighted shard sequencer + reshard internal->12 packages->3

nicktrn

committed

test(webapp): split runsReplicationService into 6 files (per-test-pg tests ~29s; keep each shard <2m)

nicktrn

committed

test: update test-timings for 6-way runsReplicationService split

nicktrn

committed

test(webapp): isolate redis for streamBatchItems to fix 30s seal-timeout flake

nicktrn

committed

perf(ci): split runsRepository 2->4 + webapp 8->10 shards (makespan ~90s)

nicktrn

committed

docs(testcontainers): document fixture choice + shard timings workflow

nicktrn

committed

test(webapp): split runsReplicationService part5 + real measured shard weights

nicktrn

committed

chore: prettier-format split test files + testcontainers package

nicktrn

committed

fix(testcontainers): author shard sequencer as CJS so config loads on node 20

nicktrn

committed

fix(testcontainers): create template_db explicitly + throw on failed schema push

nicktrn

committed

test(testcontainers): address review - drop clones, validate env, fail-fast on bad timings/push

nicktrn

committed

perf(testcontainers): offset shard bins per package so heavy files don't stack on one shard

nicktrn

committed

ci: retry docker pull in unit-test pre-pull (transient registry timeouts)

nicktrn

committed

test(schedule-engine): split scheduleEngine.test.ts so its two heavy cases shard apart

nicktrn

committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

perf(ci): speed up unit tests with sharding + container scoping#3855

perf(ci): speed up unit tests with sharding + container scoping#3855
nicktrn wants to merge 24 commits into
mainfrom
perf/speed-up-unit-tests-tri-10484

Commits on Jun 6, 2026

test(testcontainers): add local timing harness, docs, and TESTCONTAINERS_TIMING gate

perf(testcontainers): force-kill teardown + worker-scope clickhouse/redis/postgres containers

test(run-engine): make ttl consumer waits event-driven instead of fixed 13s sleeps

perf(testcontainers): worker-scope containerTest (compose pg/redis/clickhouse isolation)

perf(testcontainers): scope minio + tails; carve out replicationContainerTest for logical replication

test(webapp): move logical-replication suites to replicationContainerTest

test(run-engine): make waitpoints waits event-driven (fix the original flaky race)

test(testcontainers): add opt-in CI-resource-limit mode for local flake repro

perf(testcontainers): replicationContainerTest only needs per-test postgres (share redis+clickhouse)

fix(testcontainers): per-test redis for background-worker tests + faster redis boot

perf(ci): duration-weighted shard sequencer + reshard internal->12 packages->3

test(webapp): split runsReplicationService into 6 files (per-test-pg tests ~29s; keep each shard <2m)

test: update test-timings for 6-way runsReplicationService split

test(webapp): isolate redis for streamBatchItems to fix 30s seal-timeout flake

perf(ci): split runsRepository 2->4 + webapp 8->10 shards (makespan ~90s)

docs(testcontainers): document fixture choice + shard timings workflow

test(webapp): split runsReplicationService part5 + real measured shard weights

chore: prettier-format split test files + testcontainers package

fix(testcontainers): author shard sequencer as CJS so config loads on node 20

fix(testcontainers): create template_db explicitly + throw on failed schema push

test(testcontainers): address review - drop clones, validate env, fail-fast on bad timings/push

perf(testcontainers): offset shard bins per package so heavy files don't stack on one shard

ci: retry docker pull in unit-test pre-pull (transient registry timeouts)

test(schedule-engine): split scheduleEngine.test.ts so its two heavy cases shard apart