Comparison runs through projects. A project is a record that groups changelogs and database connections so the dashboards can show activity by application, service, or team, which makes it the dimension you slice the measurements on.

So the way you organize projects decides what you are able to compare later. A project per application lets you compare applications; a project per team lets you compare teams. See What are projects? for how to set them up.

Compare one project against another

1. Open the dashboard carrying the figures you care about.

2. Set the time range, and leave it alone for the rest of the comparison.

3. Filter to the first project and record the figures.

4. Switch the project filter and read the same figures again.

Note: The dashboards show one filtered scope at a time rather than two side by side, so a comparison means reading one scope and then the other. Holding the time range constant is what makes the two readings comparable, and Copy link to this filtered view gives you a link back to each one.

Compare a single database or changelog

For a finer grain than a project, filter to one database connection or one changelog. Coverage questions usually get answered at this level, because a project that looks well covered can still contain one connection nothing checks.

Compare environments

Environment is an attribute of a database connection rather than a filter of its own, so you compare environments by filtering to the connections that belong to each one. Every operation shows its connection’s environment as a badge, which is how you confirm you have the right set.

Cycle Time uses the same assignment to tell production from non-production, so it is worth keeping accurate for both. See Measure deployment performance.

Compare teams over time

Set each team up as its own project and you can watch it over time instead of only at one moment. Every figure carries the current window plus up to three earlier windows of the same length, so filtering to one project shows the direction that team is moving, not just where it stands. Repeating it per project is how you find where practices are diverging, and it is more useful than ranking teams against each other in a single window.