Resolved
The issue is resolved.
Monitoring
The immediate impact identified as delayed or missed scheduled function runs for a subset of customers in this cluster.
Recent metrics indicate that scheduler runtimes have returned to normal levels. To prevent recurrence, we are evaluating improvements to the service.
Investigating
Some context-api pods are unable to process function scheduler requests due to timeouts when calling Kubernetes TokenReview endpoints. The engineering team is investigating and working to restore normal schedule execution.