Feedback
Grok BotNews

·

Illustration: a corded desk telephone handset with coiled cord, in black and white.
Illustration

Grok Bot, Cloud Agents and Origin hit by multi-service outage

If a routine, cloud agent or Origin action failed between 5:15 and 7:13 PM ET on Oct 28, the outage is now resolved, so check those runs and rerun what failed.

OutageStatusAutomationsCloud AgentsOrigin

If a Grok Bot routine, cloud agent run, Bugbot review or Origin action failed Wednesday evening, it probably was not you. The Cursor status page opened a major "Multiple services degraded" incident at 5:15 PM ET and resolved it at 7:13 PM ET.

What changed#

  • 5:15 PM ET. "We are investigating service degradation for multiple services. Users may see failures across Cloud Agents, Grok Bot, and Origin. Apologies for the trouble."
  • 5:39 PM ET. Scope widened to Automations and Bugbot: "Users may see failures across Automations, Bugbot, Cloud Agents, Grok Bot, and Origin. We are working on a fix and will update shortly."
  • 6:06 PM ET. "A fix has been implemented and services are starting to recover. We will share another update shortly."
  • 6:40 PM ET. "Services have recovered and we're monitoring to make sure everything stays stable. If you're still seeing errors, please try again."
  • 7:13 PM ET. "This incident has been resolved. All services have returned to normal operation." Automations, BugBot, Cloud Agents, Grok Bot and Origin were operational.

Why it matters#

Automations was affected, so an evening digest in this window may have failed unnoticed. Cloud agent runs and Bugbot reviews started then are worth checking too.

How to use it#

  1. The status page marked the incident resolved at 7:13 PM ET, so reruns should go through now.
  2. Check each routine due between 5 PM and 7:15 PM ET.
  3. Rerun anything that failed, or use the prompt below.

Prompt (ours, untested): "List my routines that failed between 5 PM and 7:15 PM ET today and rerun each one I approve."

Good to know#

  • Staff have not posted a cause yet.
  • Two Cursor forum users report Auto-review errors around the outage, including "An error occured while classifying this action. Please review manually." One thread cites about 6:07-7:00 PM ET; another on 0.78.1 posted at 8:11 PM ET. No staff reply yet. (reported)
  • Our staff X scan since 4 PM ET found no outage post.

What to watch#

A postmortem cause, staff reply on the Auto-review reports, and whether failed scheduled runs still need a manual rerun.

Update, Oct 29: Cursor staff replied on the Auto-review reports. In one reply at 5:24 AM ET, Colin said the errors were "a separate issue that hit Auto-review itself" that lasted "until about 6:56 PM ET," with actions "reviewed normally since about 6:59 PM ET." Two of his other replies that morning (5:11 AM ET, 5:24 AM ET) described the Auto-review failure as ongoing, and the status page lists no Auto-review incident. None of the replies gives a cause for the main outage.

Sources#

Last checked: 2026-10-29

Prompt

List my routines that failed between 5 PM and 7:15 PM ET today and rerun each one I approve. (ours, untested)

Sources

  1. status.x.ai
  2. status.x.ai
  3. forum.cursor.com
  4. forum.cursor.com
  5. forum.cursor.com
  1. Known issuesGrok Bot computers are reachable again after today's outageCursor resolved the Oct 14 Grok Bot computer outage at 5:04 PM ET. Staff now say it began around 9 AM ET, and Update said files weren't ready because frozen computers couldn't be backed up.
  2. Known issuesElevated errors hitting Grok models in Cursor tonightGrok model requests in Cursor's CLI, Cloud Agents and IDE hit capacity errors from 6:25 PM ET Oct 15. Cursor marked it resolved at 1:20 AM ET Oct 16. Grok Bot stayed Operational.
  3. Known issuesCursor automations hit a platform-wide start limit for hours todayPR, Slack and webhook runs blocked by the limit were dropped, not retried. Here is how to find and re-fire them.

Something wrong? Tell the bot