Changes / Sanity / Content Agent and Canvas elevated error rate
Sanity: Major incident: Content Agent and Canvas elevated error rate
Summary
Sanity reported a major incident affecting Content Agent and Canvas: elevated error rates caused some requests to fail, though retries sometimes succeeded. It started at 2026-01-21 14:53 UTC and was resolved at 2026-01-21 15:33 UTC, lasting 40 min. The upstream AI provider implemented a fix and recovered, returning error rates to normal.
Machine-written from the diff below (Sol; type: incident degradation). Check the diff before relying on it.
What changed
Before: not present
After: Content Agent and Canvas elevated error rate
Before: not present
After: major
Before: not present
After: resolved
Before: not present
After: 2026-01-21 14:53 UTC
Before: not present
After: 2026-01-21 15:33 UTC
Before: not present
After: Canvas, Content Agent
Before: not present
After: We’re investigating elevated error rates impacting Content Agent and Canvas functionality. Some requests may fail, though retries may sometimes succeed. We’re monitoring the upstream AI provider and will share updates as we have more information.
Before: not present
After: We are continuing to investigate this issue.
Before: not present
After: We are continuing to investigate this issue.
Before: not present
After: The issue in our upstream AI provider has been identified and a fix is being implemented.
Before: not present
After: A fix has been implemented at provider level and we are monitoring the results.
Before: not present
After: This incident is resolved. Our upstream AI provider has recovered and Content Agent and Canvas error rates have returned to normal.
First version of this incident archived by Rewindex. View the live incident (opens in new tab).
History of this incident
| When (UTC) | Change | Kind |
|---|---|---|
| Major incident: Content Agent and Canvas elevated error rate (this change) Sanity reported a major incident affecting Content Agent and Canvas: elevated error rates caused some requests to fail, though retries sometimes succeeded. It started at 2026-01-21 14:53 UTC and was resolved at 2026-01-21 15:33 UTC, lasting 40 min. The upstream AI provider implemented a fix and recovered, returning error rates to normal. | Major incident |