File: docker-compose.yml:1
Severity: tech-debt
Obvious? no
replay-*, storage-*, and web use depends_on: - relay without a condition: service_healthy, and the relay service has no healthcheck: block. Compose only waits for the relay's container to start, not for the listener to bind, so workers race the relay on cold docker compose up and produce noisy reconnect storms in early logs.
Fix: add a healthcheck: to the relay (e.g. test: ["CMD", "nc", "-z", "127.0.0.1", "3340"] — see also F26 about the actual port) with sane interval/retries, then upgrade dependents to depends_on: relay: { condition: service_healthy }. Also add a healthcheck to web (curl the index) so a deploy that ships broken WASM is visible to compose-ps users.
Filed by /general-audit @ 88498a5 (2026-05-04). master: #600.
File:
docker-compose.yml:1Severity: tech-debt
Obvious? no
replay-*,storage-*, andwebusedepends_on: - relaywithout acondition: service_healthy, and the relay service has nohealthcheck:block. Compose only waits for the relay's container to start, not for the listener to bind, so workers race the relay on colddocker compose upand produce noisy reconnect storms in early logs.Fix: add a
healthcheck:to the relay (e.g.test: ["CMD", "nc", "-z", "127.0.0.1", "3340"]— see also F26 about the actual port) with sane interval/retries, then upgrade dependents todepends_on: relay: { condition: service_healthy }. Also add a healthcheck toweb(curl the index) so a deploy that ships broken WASM is visible to compose-ps users.Filed by
/general-audit@88498a5(2026-05-04). master: #600.