Activepieces had 10 outages in the last 2 years totaling 57h 33m of downtime — averaging 0.4 incidents per month.
There were 10 Activepieces outages since March 19, 2026 totaling 57h 33m of downtime. Each is summarised below — incident details, duration, and resolution information.
Worker health went down.
Workers were underperforming for couple minutes due to high load.
Worker health recovered.
Worker health went down.
Worker health recovered.
Worker health recovered.
Worker health went down.
API Health went down. no data was lost
API Health recovered.
Between 06:30 and 15:00 UTC on 2026-05-06, flow runs that made outbound HTTP calls (OpenAI, Anthropic, and most third-party pieces) failed with getaddrinfo EAI_AGAIN. The sandbox could not resolve public hostnames because its DNS servers were not on the kernel egress allowlist. A fix has been deployed and flows are running normally and all failed runs has been retried from failing step..
The issue is fixed and all failed runs has been retried from failing step..
Users that try to sign up for first time on cloud.activepieces.com via google/saml authentication were getting redirected to wrong page. which blocked them from getting into their accounts.
If you have a piece of code that calls an external API using an HTTPS URL, it may fail with the error getaddrinfo EAI_AGAIN. The change has been reverted.
The egress proxy change is now reverted
API Health went down.
API Health recovered.
Logs were experiencing significant delays (1-3 hours) in displaying. Flow execution continued normally.
Resolved. Log display delays have been fixed and logs are now showing in real-time.
Flows with delay steps experienced repeated executions due to a bug causing infinite loops, overloading Redis. All affected executions were replayed after restoration with no data loss.
Resolved. The infinite loop bug has been fixed and Redis load is back to normal. All affected executions were replayed with no data loss.
Issues arose during Kubernetes-based worker architecture migration. System entered unstable intermediate state, requiring rollback to previous deployment system.
Resolved. Rolled back to previous deployment system. All services restored to normal operation.