Skip to content

Detector Map

Quick reference for Azure Functions Diagnostics detectors and monitoring tools relevant to serverless troubleshooting.

graph TD
    A[Troubleshooting scenario] --> B{Scenario type}
    B --> C[Trigger or listener failure]
    B --> D[Performance degradation]
    B --> E[Scaling or capacity]
    B --> F[Dependency or outbound]
    C --> G[Function App Down]
    C --> H[Function Execution Errors]
    D --> I[HTTP Server Errors]
    D --> J[Function Duration]
    E --> K[Scale Controller Status]
    E --> L[Instance Count]
    F --> M[Dependency Failures]
    F --> N[DNS and Network]
    G --> O[Validate with logs and metrics]
    H --> O
    I --> O
    J --> O
    K --> O
    M --> O

How to Access

Navigate to your Function App in the Azure Portal → Diagnose and solve problems.

The diagnostics interface provides built-in detectors that analyze your function app's health without requiring custom queries.

Detector Reference

Detector Category What It Shows When to Use Related Playbook
Function App Down Availability Whether the function app is responding App not loading, health check failures Functions Not Executing
Function Execution Errors Execution Function invocation failures and error trends Rising error rates, exception storms Functions Failing
HTTP Server Errors Performance 5xx error trends for HTTP-triggered functions HTTP trigger failures High Latency
Function App Performance Performance Execution duration and latency trends Slow response investigation Timeout / Execution Limit
Scale Controller Status Scaling Scale decisions and instance allocation Queue backlog, scaling bottleneck Queue Piling Up
Linux Memory Resources Process-level memory utilization Memory pressure, OOM crashes Out of Memory / Worker Crash
Application Logs Diagnostics Function app stdout/stderr output Runtime errors, startup investigation All playbooks
Deployment Logs Configuration Deployment history and status Post-deployment failures Deployment Failures
Configuration and Management Configuration App settings and configuration validation Missing settings, runtime mismatch App Settings Misconfiguration

Azure Monitor Metrics Reference

In addition to diagnostics detectors, Azure Monitor provides metrics that can be queried via CLI or portal.

Function App Metrics

Metric Aggregation What It Shows Plan Availability
FunctionExecutionCount Total Number of function executions Y1, EP, Dedicated
FunctionExecutionUnits Total Execution units (MB-ms) Y1, EP
OnDemandFunctionExecutionCount Total On-demand executions FC1 only
OnDemandFunctionExecutionUnits Total On-demand execution units FC1 only
Requests Total HTTP requests received All plans
Http5xx Total Server error responses All plans
AverageResponseTime Average Mean response time (ms) All plans
HealthCheckStatus Average Health check probe result EP, Dedicated

CLI Quick Reference

# Function execution metrics (traditional plans)
az monitor metrics list \
  --resource "/subscriptions/$SUBSCRIPTION_ID/resourceGroups/$RG/providers/Microsoft.Web/sites/$APP_NAME" \
  --metric "FunctionExecutionCount" "Requests" "Http5xx" "AverageResponseTime" \
  --interval PT1M \
  --aggregation Total Average \
  --offset 30m \
  --output table

# Function execution metrics (Flex Consumption)
az monitor metrics list \
  --resource "/subscriptions/$SUBSCRIPTION_ID/resourceGroups/$RG/providers/Microsoft.Web/sites/$APP_NAME" \
  --metric "OnDemandFunctionExecutionCount" "OnDemandFunctionExecutionUnits" \
  --interval PT1M \
  --aggregation Total \
  --offset 30m \
  --output table

# Storage metrics (for queue-triggered functions)
az monitor metrics list \
  --resource "/subscriptions/$SUBSCRIPTION_ID/resourceGroups/$RG/providers/Microsoft.Storage/storageAccounts/$STORAGE_NAME" \
  --metric "QueueMessageCount" \
  --interval PT1M \
  --aggregation Average \
  --offset 30m \
  --output table

Application Insights Tables Reference

Table What It Contains Primary Use
requests Function invocations (HTTP triggers) Execution success/failure/latency
traces Host lifecycle, custom logs Startup, listener, runtime events
exceptions Error details with stack traces Error classification and root cause
dependencies Outbound calls to external services Dependency health and latency
customMetrics Custom-emitted and select runtime metrics Business metrics, custom counters
customEvents Custom-tracked events Application flow tracking

Detector Limitations

  • Data refresh delay: 5–15 minute lag between an event and its appearance in diagnostics.
  • Sampling: High-volume detectors may sample events rather than capturing every occurrence.
  • Consumption plan gaps: Some detectors have limited data on Consumption plan due to cold start/deallocation behavior.
  • Platform-level focus: Detectors see function execution and host events but cannot inspect application memory or stack without Application Insights.
  • Time scope: Some detectors only analyze the last 24 hours — use Log Analytics directly for older data.
  • Starting point, not conclusion: Detector output is a hypothesis generator. Always validate with logs and metrics from the KQL Query Library.

Detector vs KQL Decision Guide

Situation Use Detector Use KQL
Quick visual health check ✅
Specific time window analysis ✅
Cross-table correlation ✅
Sharing evidence with team ✅
First triage (no KQL experience) ✅
Detailed hypothesis validation ✅

See Also

Sources