Skip to content

Compare

These pages are for the question "is this capture worse than the last one?" and for the second question, "does it break our rules?". compare measures a capture against a known-good one. rules and checks hold a capture to conventions you write down. All three can fail a build.

Command The question Page
compare What got worse since the baseline: new errors, slower endpoints, bigger responses? Compare against a baseline
rules Does the capture break my rules, or a built-in rule pack? Rules and rule packs
checks Are required headers missing, or content types wrong? Built-in checks

The rules and the subsystem names live in one optional file, described in The configuration file.

A baseline is a capture you trust

compare needs a second capture to measure against: the same flow, from an earlier build, that you believe was fine. Keep it with your tests, and replace it when you accept a change on purpose. The demo capture baseline.har plays that role in this guide.

What none of them see

compare looks at errors, response time and size. It does not look at how many calls were made. A heartbeat that grew from 3 calls to 12, or an N+1, does not move it. Run diagnose on the new capture too. The recipe below does both.

Recipes