Resolved
Elevated latency and errors on document processing API - RESOLVED
Affected component: Document Processing API (partner endpoints)
Update (16:45 UTC, Nov 20): This incident is resolved. From about 16:02 to 16:36 UTC, document processing requests were much slower than usual, and some timed out or failed with HTTP 503. A security hardening change to our network infrastructure cut part of our processing unit off from external services. We resolved the issue at 16:32 UTC, and processing times were back to normal by 16:39 UTC. No data was lost. Requests that completed during the incident returned correct results. If a request failed or timed out during that window, please resubmit it. We are reviewing our security protocol change process to prevent this from happening again.
Created
Status: Investigating
Affected component: Document Processing API (partner endpoints)
Update (16:30 UTC, Nov 20): Starting at about 16:05 UTC, document processing requests have been taking much longer than normal. Some requests are timing out or failing with HTTP 503. Requests that complete return correct results. Our engineers are working on it. If a request fails or times out, please retry it with backoff. We'll post another update within 30 minutes.
Resolved