Azure Functions Health Status

Azure Functions Health Status

Use this skill to investigate the current status and health of an existing Azure Function App.

Write final answers in the user's language.

Required interaction

Ask only for missing inputs that are needed to identify the app and time window:

• Function App name (required unless already provided)
• Subscription ID or name (optional)
• Resource group (optional)
• Investigation time window, default 24h (optional)

If subscription or resource group is unknown, discover them with Resource Graph. If multiple matching apps are found, ask the user to choose one.

Fast path

1. Call Azure best practices for azurefunctions first when available.
2. Run the bundled health script for the current shell/OS:
   • scripts/get-functionapp-health-status.ps1 -AppName <app-name>
   • scripts/get-functionapp-health-status.sh -a <app-name> on macOS/Linux or Bash
   • Add -SubscriptionId <sub> and/or -ResourceGroup <rg> when known.
   • Add -Hours <n> when the user requests a specific time window.
   • For Bash, use -s <sub>, -g <resource-group>, and/or -H <hours> when known.
   • Bash script requires Azure CLI plus a working python3 or python executable for JSON shaping.
3. If bundled script execution is unavailable, use references/health-status-commands-and-kql.md.
4. Report health findings and action items; avoid dumping raw logs unless requested.

If health findings require runtime or trigger-specific interpretation, use ../azure-functions-common/references/routing.md on demand. Do not load shared language/extension references for plain status output.

Scope boundary

This skill should collect:

• Current app state: enabled/running/stopped, metadata availability, runtime availability
• Plan status and Resource Health result
• Azure Monitor metrics: instance count, CPU, memory, execution counts/units, always-ready units
• Trigger status summary and notable disabled or indexing-failed functions
• Application Insights/Log Analytics: request counts/failures/p95, dependency failures, exceptions, warning/error traces
• Recent Activity Log events
• Gaps: missing telemetry, unsupported Resource Health, unavailable diagnostic data

Do not perform full static inventory beyond fields needed to interpret health. Use the inventory skill for detailed specifications.

Next steps

• On success, suggest azure-functions-diagnostics to interpret findings and recommend fixes when health signals are degraded or unclear.
• On failure, suggest azure-functions-inventory to confirm app identity and configuration before retrying health checks.

Interpretation rules

• For Flex Consumption, Resource Health may return Unknown or Unsupported; rely on app metadata, plan status, metrics, telemetry, and Activity Log.
• Workspace-based Application Insights uses AppRequests, AppExceptions, AppTraces, and AppDependencies; classic tables may be empty.
• Treat unresolved %SETTING_NAME% placeholders in traces/exceptions as configuration/indexing issues.
• Redact connection strings, keys, tokens, and storage secrets. It is safe to report setting names and whether values exist.

Output template

Target: <app> (<resource-group>, <subscription>, <region>)
Current status: app <state>, metadata <availability/runtime>, plan <status>, Resource Health <state/reason>
Metrics: instances <max/avg>, CPU <max>, memory <max>, executions <summary>
Execution: requests <n>, failed <n>, p95 highlights <...>; dependency failures <n>
Triggers: enabled <n>, disabled <n>; indexing/config issues <summary>
Recent changes: <Activity Log summary>
Findings: <1-3 bullets>
Recommended next actions: <1-3 bullets>
Gaps: <unavailable telemetry or unsupported checks>