Skip to content
All outages

Editor's note: Analogue: a quota service behaves like a Redis session store.

Incident · Medium · Redis · Infra

Zero usage, zero quota

  • SEV1
  • GLOBAL
  • ~47 MIN
90s

The situation

Nobody can sign in. Every product that needs a logged-in user fails worldwide for about 45 minutes. The identity service's data is intact and nothing was deployed to it today. Your internal support tools need sign-in too, so they are down as well.

The evidence

Identity store: reported usage vs granted quotaillustrative
  • reported usage
  • quota
Logs and signalsReconstructed
Oct quota migration: old system reports usage = 0
Oct grace period started
14 Decgrace period expired → automation lowers storage quota
+1m identity store leader: write rejected, over quota
+2m auth lookups: data too stale, refusing to serve

Your call

What took sign-in down?

Charts and logs are reconstructions. Aim for 90 seconds; going over costs nothing.

Last updated: Official postmortem