One event loop: the thread error is gone
357 thread errors in the outage · 0 in 3,375 attempts after
Engineering
Production notes
Incidents from the worker that refreshes visa-sponsored European software jobs every six hours. Numbers from production. No runbooks.
can't start new thread: 186 companies looked dead
186 / 394 companies flagged · 357 / 358 errors were the same RuntimeError
678 Postgres round-trips: the board was not hung
678 DB round-trips for 98 companies · Germany ~89s → ~1.7s
Cache check in the hot path: Redis did what Postgres did
2 Redis round-trips × N companies · board back to ~1.3–1.7s
Playwright hung for 15 hours and the worker still looked Up
97/98 companies done · scheduler blocked 15+ hours · no join timeout