Workers spawned but stuck at 'Starting...' (Passenger killed each before it could reach 'Ready'). Caused by rapid push cadence accumulating zombie next-server processes under one HelperAgent. Recovered with a single mass-kill + Passenger restart per the §5 runbook within 30 seconds. ~10 minute disruption.
Overall: Operational
Last checked: 2026-10-18 15:47:59 UTC
Web app (pdfcraftai.com)
If you're reading this page, the web app is up. (Manual probe — no automated check.)
Sign-in & accounts
NextAuth (Credentials + Google) — depends on the database.
Live DB ping 16ms
Free PDF tools
Browser-side processing — never touches our infrastructure for the file bytes themselves.
AI tools
Chat, summarize, translate, OCR, redact — depends on configured AI providers.
3 providers configured
Bulk processor
Zip / folder drops with manifest output. (Manually flagged.)
Public API
REST endpoints + webhooks. (Manually flagged.)
RECENT INCIDENTS
Last 30 days.
503 during deploy — accumulated next-server zombies
2026-05-08 · RESOLVEDStale-worker hold after 175 + 192 deploys
2026-04-28 · RESOLVEDAuto-pull completed but the runtime kept serving the prior commit because next-server processes held stale code in memory. Cleared with the documented SSH pkick recipe; deploys now flip cleanly.
Razorpay checkout failed with config error
2026-04-22 · RESOLVEDProperty-name drift between adapter (writing keyId:) and checkout button (reading .key) caused 'Authentication key was missing' on every Buy click. Fixed in commit cb41e14 + new contract pin in test-razorpay-handoff.mjs.
Sign-in errors after IPv6 cutover
2026-04-19 · RESOLVEDMySQL connection pool had IPv6 literals in the URL. Fixed by switching to per-component env vars and restarting the node runtime. Duration: ~35 minutes.
Something looks off?
Write to support@pdfcraftai.com — we'll investigate and post an update here.
Report a problem