Din't find any errors in logs. Tracking down the logs by tracing ID, it gives multiple calls of Lifecycle endpoint (for some reason the same tracing ID is being used multiple times).
First request has been processed in 255 ms.
However, next call (which supposed to be the same) has been stuck for 73 s (which may have caused the timeout in the Pub client).
Lifecycle metrics
During that time, there was a noticeable spike of active requests and active threads, while maintaining the ordinary number of requests per second and CPU usage.
(issue originally reported by Anna)
It looks like a performance issue in Lifecycle.
Job logs
Lifecycle logs
Din't find any errors in logs. Tracking down the logs by tracing ID, it gives multiple calls of Lifecycle endpoint (for some reason the same tracing ID is being used multiple times).
First request has been processed in 255 ms. However, next call (which supposed to be the same) has been stuck for 73 s (which may have caused the timeout in the Pub client).
Lifecycle metrics
During that time, there was a noticeable spike of active requests and active threads, while maintaining the ordinary number of requests per second and CPU usage.