Fixing High CPU and Memory Usage in Azure App Service

Fixing High CPU and Memory Usage in Azure App Service

Your app is running fine, and then suddenly… users complain about slowness or the app even crashes.
When you check the Azure Portal, you see CPU pegged at 90–100% or memory exhausted.
This is a very common real-world issue in Azure App Service. Let’s walk step by step.

Step 1: Monitor Metrics

Go to: Azure Portal → App Service → Metrics → CPU Percentage / Memory Working Set Look for spikes during:

• High user traffic
• Background jobs
• Large API calls

You can also run via CLI:

az monitor metrics list \
  --resource /subscriptions/<subid>/resourceGroups/MyRG/providers/Microsoft.Web/sites/MyAppService \
  --metric "CpuPercentage" --interval PT1M

This gives you CPU usage minute by minute.

Step 2: Use App Service Diagnostics

• In the App Service → Diagnose and Solve Problems:
• Select Availability & Performance
• Check High CPU Analysis or High Memory Analysis This shows you:
• Top requests consuming resources
• Long-running threads
• Memory leaks

Step 3: Profile the App

• Enable Application Insights Profiler:
• Portal → App Service → Application Insights → Turn On Profiler
• Capture snapshots during high load
• Review flame charts to see slow functions or memory leaks

Step 4: Scale Out or Scale Up

• Sometimes, the fix is simple: add more resources.
• Scale Up → Increase plan (B1 → S1 → P1v3)
• Scale Out → Add more instances

az appservice plan update \
   --name MyAppPlan \
   --resource-group MyRG \
   --number-of-workers 3

This adds 3 worker instances to balance load.

Step 5: Optimize the App

• High CPU isn’t always Azure’s fault. Check for:
• Infinite loops in code
• Heavy synchronous operations
• Poor database queries

✅ Use async programming ✅ Cache results in Azure Cache for Redis ✅ Optimize database queries with indexes

Step 6: Use Autoscale

• Enable Autoscale Rules to handle traffic spikes:
• Portal → App Service Plan → Scale Out (App Service Plan)

Example:

• Scale out if CPU > 75% for 10 minutes
• Scale back in when CPU < 40%

Step 7: Restart and Recover

If your app gets stuck:

az webapp restart --name MyAppService --resource-group MyRG

Sometimes a simple restart clears memory leaks temporarily until you apply a permanent fix.

Common Causes of High CPU/Memory

• Unoptimized Code → Loops, blocking threads, recursive calls
• Leaky Dependencies → Not disposing connections properly
• Bad DB Queries → Full table scans
• Large Payloads → JSON/XML parsing at scale
• Traffic Surge → Sudden load without autoscale

Best Practices

• Always enable Application Insights for monitoring
• Use Azure Front Door or Azure CDN to offload static content
• Offload long jobs to Azure Functions or Azure WebJobs
• Test performance with Azure Load Testing

Final Take

High CPU and memory issues in Azure App Service are often a mix of code inefficiency + resource limits. With monitoring, scaling, and profiling tools, you can quickly identify whether the fix is your app or your plan.

Remember: Don’t guess → Measure → Fix → Scale.

– Kasi @ KasdevTech or LinkedIn