Response caching and usage quotas

Azure API Management

Fiodar Sazanavets

Senior Software Engineer at Microsoft

Setting up response caching

  • When responses are cached, there's no need to go to the backend endpoint
  • Caching is applied to popular endpoints that often return the same data
  • Round-trips to endpoints are time consuming and expensive
  • Can be implemented in APIM via caching policies

Response caching

Azure API Management

Configuring response caching

  • Caching policies configuration can:
    • Specify cache duration
    • Vary by query string or headers
    • Set the location of the cache
  • Cache location can be:
    • The edge, i.e. API gateway
    • External caching service, e.g. Redis

Caching policies

Azure API Management

Protecting against request surges

Rate limiting

  • Even with caching, too many clients can overload APIs
    • Causes slower performance
    • Increases resource costs
    • Worst case: system outage (e.g., DDoS attack)
Azure API Management

How usage quotas protect APIs

  • Restrict number of calls over a period
    • This guards against sudden spikes of requests
  • Can protect against DDoS attacks
  • Can save money by preventing overuse

Azure API Management

Measuring resource usage

Application Insights

  • You can't improve what you don't measure

    • Enable Application Insights to track metrics (e.g., latency, outliers)

  • Example: 'GET Products' call

    • Low cache hit ratio but high traffic
    • Solution: increase cache duration (5 -> 30 minutes)

  • With insights, you can fine-tune:

    • Cache durations
    • Quota thresholds
    • Service tiers for best cost-performance
Azure API Management

Let's practice!

Azure API Management

Preparing Video For Download...