chore(tradein/deploy): don't recreate scraper on infra-only deploys + 120s graceful stop (phase 0+1)
All checks were successful
CI / changes (pull_request) Successful in 9s
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped

The tradein-scraper container runs the same image as backend
(python -m app.scheduler_main, in-app scheduler) and owns the in-flight
scrape sweeps (a browser card takes ~15-27s). Two problems compounded to
kill running jobs on every deploy:

Phase 0 (deploy-tradein.yml): SCRAPER_CHANGED also fired on the `infra`
paths-filter (compose / workflow / deploy/**), so any generic infra edit
recreated the scraper container and SIGKILLed the running job. Dropped the
`|| infra == 'true'` term — only real scraper-code paths (already covered
by the `scraper` paths-filter) or a manual workflow_dispatch recreate it.

Phase 1 (docker-compose.prod.yml): the scraper had no stop_grace_period,
so Docker's default 10s window SIGKILLed an in-flight card (15-27s) before
it could finish. Added stop_grace_period: 120s + explicit stop_signal:
SIGTERM so a running unit can finish + checkpoint. The cooperative-drain
handler that consumes SIGTERM lands in a later phase; this is the
foundation.
This commit is contained in:
bot-backend 2026-06-28 19:30:14 +03:00
parent 57ae2dfe00
commit 8aaa42f101
2 changed files with 12 additions and 1 deletions

View file

@ -276,7 +276,13 @@ jobs:
env:
IMAGE_TAG: latest
GHCR_PAT: ${{ secrets.GHCR_PAT }}
SCRAPER_CHANGED: ${{ needs.changes.outputs.scraper == 'true' || needs.changes.outputs.infra == 'true' || github.event_name == 'workflow_dispatch' }}
# Phase 0: generic infra edits (compose / workflow / deploy/**) must NOT
# recreate the scraper and SIGKILL a running multi-hour job. Only genuine
# scraper-code paths (the `scraper` paths-filter already covers
# app/services/scrapers/**, scrape_pipeline.py, scheduler.py,
# scheduler_main.py, app/tasks/**) — or a manual workflow_dispatch —
# should trigger a scraper recreate. (infra term intentionally dropped.)
SCRAPER_CHANGED: ${{ needs.changes.outputs.scraper == 'true' || github.event_name == 'workflow_dispatch' }}
GITHUB_SHA: ${{ github.sha }}
with:
host: ${{ secrets.DEPLOY_HOST }}

View file

@ -129,6 +129,11 @@ services:
browser:
condition: service_started
restart: unless-stopped
# Phase 1: give the in-flight scrape unit (browser card ~15-27s) time to finish
# + checkpoint on SIGTERM before Docker SIGKILLs. Default grace is only 10s, which
# SIGKILLs a running card mid-flight. (Cooperative-drain handler lands in a later phase.)
stop_grace_period: 120s
stop_signal: SIGTERM
networks:
- tradein-net