Replica Crash Signals¶

Use this query to identify restart-heavy replicas and crash-loop patterns by revision.

Data Source¶

Query Pipeline¶

flowchart LR
    A[Filter by app] --> B[Filter crash and restart terms] --> C[Summarize by replica and revision] --> D[Sort by event count]

Query¶

let AppName = "my-container-app";
ContainerAppSystemLogs_CL
| where ContainerAppName_s == AppName
| where Log_s has_any ("CrashLoopBackOff", "terminated", "exited", "restart", "OOM")
| summarize events=count(), firstSeen=min(TimeGenerated), lastSeen=max(TimeGenerated) by RevisionName_s, ReplicaName_s
| order by events desc

Example Output¶

Interpretation Notes¶

• High events on few replicas indicates unstable runtime behavior.
• OOM-related terms suggest memory pressure rather than logic-only defects.
• Normal pattern: low restart count outside deployments.

Limitations¶

• Event detail quality varies by platform event source.
• Requires correlation with console logs for stack traces.

See Also¶

• Latest Errors and Exceptions
• CrashLoop OOM and Resource Pressure Playbook