Is your feature request related to a problem? Please describe.

When testing a new Mimir installation, it would be nice to perform a simple smoke test to make sure everything is wired up as expected.

Describe the solution you'd like

The mimir-continuous-test tool could be extended to support a second operating mode which runs the embedded tests as a smoke test.

Because the system may take some time to become healthy on initial install, it would be helpful to support running the tests multiple times until they succeed (up to some limit). Alternatively, we could rely on external automation to re-run the tests until success. For example k8s Job objects can be configured to retry on failure.

The process exit status should also change to reflect test success / failure when running in smoke test mode. This will make it possible to use the tool in automation (e.g. helm test).

Describe alternatives you've considered

I could use an existing tool like benchtool to validate system functionality, but mimir-continuous-test is designed to be extended with additional tests (e.g. Ruler API) over time. These new tests will be very useful in a smoke testing context as well.

Additional context