Glean had 26 outages in the last 2 years — averaging 1.1 incidents per month.
There were 26 Glean outages since April 9, 2025. Each is summarised below — incident details, duration, and resolution information.
We are currently seeing elevated chat errors affecting our customers hosted on AWS due to availability issues with Amazon Bedrock. The impact is primarily affecting requests using Claude Opus 4.6 and Claude Sonnet 5. Our team is actively monitoring the situation.
Some users may experience authentication failures or errors when using extension action packs (e.g., Jira, Confluence, Outlook, Teams). Users attempting to authorize or use these extensions may see HTTP 401 or 500 errors. Our engineering team is actively deploying fixes to affected environments. We expect recovery to progress incrementally as the mitigation completes across deployments. We will continue to closely monitor system behavior as the mitigation completes.
Some users are experiencing authentication failures or errors when using extension action packs (e.g., Jira, Confluence, Outlook, Teams). Users attempting to authorize or use these extensions might have seen HTTP 401 or 500 errors.
Some deployments had issues with executing some of the actions, we have mitigated the issue for almost all the customers
We are rolling out a mitigation for the issue affecting Atlassian and other related actions. Impact now: Fix deployed to most of the users. Some users may still experience with Atlassian and related actions, but we expect it to recover as the mitigation completes. We will continue to closely monitor system behavior as the mitigation completes.
We have deployed a fix for the issue affecting extension actions, and early signals show error rates have reduced and things are back to normal. We are keeping the systems under heightened monitoring to ensure they remain stable. Impact: Most users should now be able to use the extension actions. If you continue to experience issues, please contact Glean support.
We are rolling out a mitigation for the issue affecting chat and assitant. Impact now: we expect improving as the mitigation completes. We will continue to closely monitor system behavior as the mitigation completes.
We have disabled the change and confirmed that user experience has returned to normal.
The incident has been resolved. We apologize for the disruption and appreciate your patience while we addressed this issue.
We have identified the cause of the issue. We are currently implementing a fix and monitoring its effect.
Incident: Client Token management API is failing We have identified the cause of the issue. We are currently implementing a fix and monitoring its effect.
We have deployed a fix for the issue globally. We are keeping the systems under heightened monitoring to ensure they remain stable. Impact: Most admin users should now see the Client Token management functionality restored. If you continue to experience issues, please contact Glean support.
The incident has been resolved. We apologize for the disruption and appreciate your patience while we addressed this issue.
We have identified elevated rate limiting from Google Vertex AI. Leading to this error on Glean - "Vertex AI is at capacity right now, so I can't get a response. Give it a few minutes and try again." Our engineering team continues to work with Google support on this issue. Symptom Affected customers may encounter transient errors (429/529) when making requests to the Claude Opus 4.6 model via Glean Chat or Agents. Workaround Affected customers may consider using alternative models.
We are experiencing an issue with Vertex AI Online Prediction beginning on Sunday, 2026-05-17 21:10 PDT. Our engineering team continues to investigate the issue. We will provide an update by Monday, 2026-05-18 11:00 PDT with current details. We apologize to all who are affected by the disruption. Symptom Affected customers may encounter transient errors (429/529) when making requests to the Claude Opus 4.6 model Workaround Affected customers may consider using alternative models.
We have identified elevated rate limiting from Google Vertex AI. We are monitoring the issue and will share updates as service recovers.
GCP has resolved the issue with Vertex AI Online Prediction for all affected users as of Monday, 2026-05-18 07:49 PDT. We thank you for your patience while we worked on resolving the issue.
The incident has been resolved. Impact window: From May 18, 2026, 9:00 AM UTC to Monday, May 18, 2026, at 14:49 UTC, some users experienced elevated rate limiting from Google Vertex AI. Remediation: The issue with Vertex AI Online Prediction has been resolved for all affected users. We apologize for the disruption and appreciate your patience while we addressed this issue.
The incident was triggered by an AWS availability issue affecting underlying compute and storage in a single availability zone. That event left the search cluster without a healthy master node, which caused search requests to time out or fail. During recovery, the master node could not restart because its persistent volume was still marked in use on the impacted instance, delaying failover until the volume attachment problem was resolved and healthy search capacity was restored. This is affecting only the AWS hosted instances. Please refer AWS status page: https://health.aws.amazon.com/health/status?path=open-issues
We have successfully restored all Glean instances in the us-east-1 region that were impacted by the recent AWS EC2 outage.
We are currently observing increased latencies and intermittent server errors with respect to Bedrock models. Our team is actively monitoring the issue .
The incident has been resolved. Impact window: From 1:00 AM to 4:15 AM PDT, some users experienced degraded AI search and assistant performance, including slower response times, due to elevated latency in AWS Bedrock's US-WEST-2 region. Remediation: AWS rebalanced capacity and applied scaling adjustments to their infrastructure, and we have confirmed that user experience has returned to normal. We apologize for the disruption and appreciate your patience while the issue was addressed.
We have rolled out a permanent fix to this issue
We finished rolling out the fix.
The issue is resolved.
The issue of authentication errors for Jira and Confluence is now resolved. The root cause was an IP block from Atlassian which was triggered by a spike in the number of user token requests. We have worked with Atlassian to get the IP block removed. We have also made fixes to fail gracefully when Atlassian starts returning errors for auth requests.
Google has confirmed multiple GCP products are affected in us-central1. Multiple cloud services impacted Issue has been mitigated now after a confirmation from google
Fix has been rolled out successfully and users are no longer facing login issues
Google confirms the issue is fixed.
The issue has now been verified fixed.
Changes have been rolled back on affected customer.
The issue has been resolved.
We are not observing any more errors
The issue was fixed at approximately 5:30 PM IST today.After resolution, we've been monitoring & confirm that all services/deployments are operating normally. Our team is conducting a comprehensive root cause analysis and will implement any necessary proactive measures to help prevent similar incidents in the future. Thank you for your patience and understanding.
Azure metrics are lookin stable now.
GCP issue is resolved
Beginning 4:20pm PT we see the Azure OpenAI availability to be recovered.
The original issue is now resolved. A latent issue related to the admin console that affected a few customers was identified and resolved quickly. We thank you all for your patience through this major incident.
All regions have recovered
Google Infrastructure team was executing a routine maintenance action to shut down an unused Virtual Private Cloud (VPC) controller in a single Google Cloud zone. Unfortunately, a bug in the automation caused the component to be shut down in all zones where it was still in use which affected Glean's functionality. Google product teams have restored the services now - please refer Google status page for more details.