Monitoring metric set size

[kelunik]

Is your feature request related to a problem?

We're running into the default metric_set_limit (#2148, #2717) (probably due to many different transaction names). We can obviously increase the limit now based on a good guess, but it doesn't seem possible to examine the current metric set size.

We'd like to be able to monitor the metric set size to ensure we're well below the limit, but still keep the limit rather low to prevent explosion.

Describe the solution you'd like

It'd be great if we could add a micrometer gauge that monitors the metric set size, or if this becomes a built-in metric.

Describe alternatives you've considered

Raising the limit higher than necessary.

[jackshirazi]

would periodically logging the value work for you?

[kelunik]

Logging would be a good start, but my preferred way would be exposing the size either via the health metrics or via a public API, so I can create my own metric from that.

[jackshirazi]

Adding a DEBUG log is easily done. Adding anything else is not going to be prioritized by the Elastic team as there are not a lot of people requesting this, it's quite niche. So I can add a DEBUG log, or I'm happy for you to create a PR with your proposed addition!