Astronomer change history

Astro UI and API appeared offline for some users in a major incident starting 2026-10-13 18:02 UTC, resolved at 2026-10-13 19:53 UTC (1 h 51 min). Existing tasks and Dag Runs continued successfully. Restarting a service experiencing higher-than-expected load restored functionality; Astronomer identified a proximate cause but did not disclose it.

Astronomer reported a critical incident involving tasks stuck queued or failing on several hosted deployments, initially citing worker autoscaler scaling issues. It started at 2026-10-01 22:40 UTC and was resolved at 2026-10-01 23:47 UTC, lasting 1 h 7 min; the final update called it a false alert and said multiple deployments were not affected.

The Astro Cloud UI Deployments view was unavailable in a major-impact outage starting 2026-09-24 14:52 UTC, lasting 41 min and resolved at 2026-09-24 15:34 UTC; task execution was unaffected. Astronomer rolled back an identified breaking change and advised customers still experiencing issues to refresh the page, then open a support ticket if needed.

Astronomer reported a major outage from 2026-09-17 17:54 UTC to 2026-09-17 21:50 UTC, lasting 3 h 56 min. A regression from a change rolled out that day silently dropped custom-role permissions, causing reported deploy failures for tokens using those roles; re-adding deployment.deploys.create mitigated failures. All custom roles were restored, and the incident was resolved.

Astronomer reported a major Astro degradation: tasks remained queued on some worker queues because autoscaling resource names exceeding Kubernetes’ 63-character limit were rejected, preventing worker scaling for a very small number of customers. It started at 2026-09-08 12:33 UTC and resolved at 2026-09-08 20:09 UTC, lasting 7 h 35 min; a fix was implemented.

Astronomer reported a major outage affecting authentication, deployment access and DAG visibility, with 401 and RBAC errors for multiple customers. It started at 2026-09-04 03:08 UTC and was resolved at 2026-09-04 06:07 UTC, lasting 2 h 58 min. A network configuration change during an infrastructure upgrade disrupted control-plane connectivity to customer deployments; Astronomer applied a fix and confirmed normal operation.

Astronomer reported a major degradation from 2026-08-18 18:51 UTC to 2026-09-10 22:19 UTC, lasting 23 days 3 h. A metrics reporting bug falsely flagged missed scheduler heartbeats, affecting Astro UI dashboard and monitoring visibility; task execution, scheduling, and workflow processing were unaffected. Astronomer reported rolling out a fix and later marked the issue resolved.

Astro Hosted experienced intermittent task instance failures with major impact from 2026-08-18 04:17 UTC to 2026-08-18 07:36 UTC, lasting 3 h 18 min. Astronomer identified an issue affecting an internal rollout related to the pre_execute Airflow hook and resolved the incident after completing a fix rollout across all Astro deployments.