Video: Challenges of API Testing


Cloud native API Testing comes with a lot of challenges. In this video see how to overcome these challenges with the novel approach of traffic replay.

Challenges of Cloud Native API Testing

  • How do clients call the API?
  • Required dependencies can be missing.
  • Test data has to be reset in between runs.

What is the Ideal API Testing Method?

  • Identify what needs to be tested.
  • Generate automation quickly.
  • Run in a stable environment with the right data.

Identify what needs to be tested

  • See the exact set of calls.
  • Identify the technologies used by the API.
  • Drill down to deep detail like request and response headers and payloads.

Generate API Test automation quickly

  • Turn API traffic into test cases.
  • Update security tokens, parameters, timestamps, etc.
  • Check responses to every call.

Run in a stable environment

  • Detect microservice dependencies.
  • Turn API traffic into service mocks.
  • Orchestrate the data and environment.

Use Case: Performance Testing for SRE

  • Track the SRE Golden Signals:
    • Latency: How long the API takes to respond.
    • Throughput: How many transactions per second can the API handle.
    • Success Rate: How many of the calls returned the correct value.
    • Saturation: How did the underlying infrastructure perform under load.

Sre golden signals

Use Case: Functional Testing for SDET

  • Assertions for every single call:
    • Compare the actual against expected values.
    • Visually demonstrate the differences.
    • Integrate with CI to shift-left testing at code check-in or merge.

Sdet functional

Value of Speedscale

  • Software Engineers
    • Spend less time writing scripts and more time building new features.
    • Take out the guesswork by building realistic scenarios from live application traffic.
  • SRE / DevOps
    • The CI system runs faster without full E2E environments with lots of moving parts.
    • Automatically produce SLO numbers to ensure the API meets expectations.
    • Perform more validations to find issues before they reach production.
  • Executives
    • Deployments are safer when more automated testing is performed before each release.
    • Cloud spend for non-prod can be up to 40% of the bill, reduce it with smaller environments.

Get started for free

ProxyMock desktop or Speedscale Cloud — choose your path.