CI/CD Platforms

Overview

Integrate TestKase reporting into your CI/CD pipeline to automatically push test results after every build. Each platform guide provides a ready-to-use pipeline template with the @testkase/reporter CLI.

Supported Platforms

Platform	Guide
GitHub Actions	GitHub Actions Guide
GitLab CI	GitLab CI Guide
Jenkins	Jenkins Guide
Azure DevOps	Azure DevOps Guide
CircleCI	CircleCI Guide

Common Pattern

Every CI/CD integration follows the same three-step pattern:

Store secrets — Add TESTKASE_PAT as a CI/CD secret or environment variable. Never hardcode tokens in pipeline files.
Run tests with continue-on-error — Your test step should not block the pipeline on failure. Use continue-on-error: true (GitHub Actions), allow_failure: true (GitLab CI), or equivalent.
Run the reporter with always-execute — The reporting step must run regardless of test pass/fail. Use if: always() (GitHub Actions), when: always (GitLab CI), or post { always { } } (Jenkins).

Example Reporter Step

npx @testkase/reporter report \
  --token $TESTKASE_PAT \
  --project-id PRJ-1 \
  --org-id 1173 \
  --cycle-id TCYCLE-5 \
  --format junit \
  --results-file test-results/junit.xml \
  --build-id "$BUILD_NUMBER"

--cycle-id TCYCLE-5 above is a placeholder. The test cases you report on must already be linked to the cycle you target. If they are not, the API rejects every result with Test case not linked with this test cycle, the reporter exits 1, and this pipeline goes red on its first run. Either add those test cases to the cycle first, or drop --cycle-id altogether and report into TCYCLE-1 — the Master Cycle, which always contains every test case in the project.

Always include --build-id with your CI build number. This links each execution to a specific build, making it easy to trace results back to code changes.

Branching on the exit code

The reporter uses a different exit code per failure class, so a pipeline can distinguish a reporting problem from ordinary rejected results:

Code	Meaning
`0`	Reported successfully
`1`	Some results were rejected by the API — usually test cases not linked to the target cycle
`2`	Authentication, network or API failure
`3`	Results file missing, or the wrong `--format` for the file
`4`	Untagged or unknown Automation IDs, with a `fail` policy set

npx @testkase/reporter@1 report ... || code=$?
if [ "${code:-0}" -ge 2 ]; then
  echo "TestKase reporting itself failed (exit $code)"
  exit "$code"
fi

Full detail in the Exit Codes reference.

A 0 does not prove results arrived. The reporter also exits 0 when the results file parsed but contained no tests, when no Automation IDs matched the pattern, and when every attachment upload failed. Read the Reported: N succeeded, N failed summary line in the job log on your first run.

Next Steps

Choose your platform guide above and follow the step-by-step instructions.
See Best Practices & FAQ for troubleshooting and optimization tips.
See CLI Reference for the full list of reporter flags.