Monitoring as Code: Learn more about the API Check Construct.
Light mode interface

​
What are API Checks?

API checks are automated HTTP requests that validate API endpoint functionality by checking status codes, response content, headers, and performance metrics. They simulate real API usage patterns and help catch issues before they impact your users. You can create API checks in the UI, via the API Check Construct or by importing your existing Swagger/OpenAPI spec. API checks are perfect for:
  • REST API endpoint monitoring
  • GraphQL API validation
  • Authentication system testing
  • Third-party API dependency monitoring
  • Payment and transaction APIs
  • HTTP method support (GET, POST, PUT, DELETE, etc.)
  • Request/response validation
  • JSON schema validation
  • Authentication handling
  • Performance measurement
  • Global monitoring locations
If you are looking to monitor complex API flows or data processing, you might want to consider using Multistep Checks.

​
How API Checks Work

API checks perform comprehensive endpoint validation:
  1. HTTP Request - Sends configured requests with headers, body, and authentication
  2. Response Analysis - Captures status codes, headers, body content, and timing
  3. Validation - Runs assertions against response data and performance metrics
  4. Performance Tracking - Measures response times and identifies trends
  5. Alert Generation - Notifies teams when endpoints fail or degrade
API checks are ideal for testing both internal services and external API dependencies your application relies on.

​
Timeouts

All API checks are capped at a timeout of 30 seconds. With each request, we record the most relevant timing phases. This can help you troubleshoot slow responses, e.g. your DNS might be slow. The timing phases correspond to the Node.js request library timing phases:
  • wait: Duration of socket initialization
  • dns: Duration of DNS lookup
  • tcp: Duration of TCP connection
  • firstByte: Duration of HTTP server response
  • download: Duration of HTTP download
api monitoring timing phases

​
Interpreting the wait metric

The “wait” time is the time it takes for the underlying Node HTTP module to get a socket from the network layer. This time can vary quite a bit, usually it goes from 900 microseconds to some milliseconds. Most clients for API / webpage will have this lag too.