This page is written by the engineer on call and updated by hand when something changes. It is hosted separately from the image network.
All systems operationalLast updated 18 Nov 2026, 09:40 UTC
Availability 21 Aug – 18 Nov 2026, one bar per day
Image delivery100.00%
Transform workers99.98%
Origin fetch99.96%
Purge API99.99%
Dashboard99.93%
Incident history most recent first
Purge requests delayed resolved · 41 min
Between 14:12 and 14:53 UTC purge requests were accepted but applied up to 30 minutes late. A queue consumer had stalled after a routine deploy. Image delivery was not affected. We now alert on queue age rather than queue length.
Slow cache misses in asia-2 degraded · 1 h 26 min
A storage node in asia-2 was replaced and the remaining nodes rebuilt under load. Cached images were served normally; new variants took 300–900 ms to render. Traffic for uncached variants was moved to asia-1 at 07:31 UTC and moved back the next morning.
Origin fetch timeouts for S3 origins in us-east-1 degraded · 2 h 05 min
An upstream networking problem at the storage provider caused about 4% of first-time fetches to time out with 504 origin_timeout. Images already known to any region were served from cache. No action was needed from customers.
Dashboard unavailable resolved · 18 min
An expired internal certificate took the dashboard offline from 05:02 to 05:20 UTC. Image delivery and the API were unaffected. The certificate is now renewed automatically like the public ones.