Knowledgebase

Testing in Continuous Delivery Print

  • softwaretestingqa, software, performance, troubleshooting, security, guide, howto, solution
  • 0

Verification in a pipeline.

WHAT THE PIPELINE SHOULD DO

Build Run fast tests Run static analysis and dependency scanning Run integration tests Deploy to a test environment Run end-to-end tests Deploy onward, on success

WHAT ORDER MATTERS FOR

Failing fast, so feedback is quick.

WHAT TO RUN FIRST

Whatever is fastest and most likely to fail.

WHAT DURATION TO TARGET

Feedback on a change within minutes.

WHY

Beyond that, developers move on and context is lost.

WHAT TO DO IF THE SUITE IS TOO SLOW

Parallelise Move tests down the pyramid Run slower suites separately, on a schedule

WHAT A BROKEN PIPELINE MEANS

Nobody can release.

WHAT THAT DEMANDS

Fixing it immediately, above other work.

WHAT TO NEVER DO

Disable failing tests to unblock the pipeline.

WHY

The next real failure is then invisible.

WHAT GATES TO ENFORCE

Tests passing Coverage not falling substantially No new high-severity security findings

WHAT TO AVOID GATING ON

Metrics that produce gaming rather than quality.

WHAT TO MONITOR

Pipeline duration Failure rate Time to fix a broken pipeline


Was this answer helpful?
Back

Are you happy with your experience? Leave us a review on Trustpilot.


Trustpilot