Lesson225 words

Monitoring pipeline health

Monitor pipeline health

The syllabus names three areas without ranking them. Measure each with a defined report rather than a subjective proxy.

AreaDefined measure and report use
Failure/pass rateSuccessful or failed runs divided by total runs over a chosen period; outcome reports also expose failed-run count by pipeline
DurationHow long successful or partially successful runs take, trended by day and percentile
Flaky testsTests with differing outcomes despite unchanged source/execution environment; report count/rate and affected runs

Flaky-test behavior

Azure DevOps Services system detection can rerun a failed test in the same pipeline execution; a fail-then-pass result can be marked flaky. Custom detection can also mark or unmark tests through the API. Flaky result data and tags support reports and troubleshooting. Suppressing flaky failures from the test summary or pipeline result is a configurable choice, not an automatic rule.

An individual run remains useful for diagnosing that execution. Trends add a different view: they reveal changes and recurring patterns in pass/fail rate, completion duration, test duration, failures, and non-deterministic outcomes. Compare like-for-like periods and investigate the underlying runs rather than treating a trend alone as a diagnosis.

Primary sources

Ready to study Designing and Implementing Microsoft DevOps Solutions (AZ-400)?

Practice tests, flashcards, and all study notes — free, no sign-up needed.

Start Studying — Free