Context

I'm trying to figure out if the Sampling Override preview feature can be used to reduce sampling on metrics.

The doc here doesn't help unless I want to work on HTTP or JDBC calls.

We are using Kafka and while we get all the client-side metrics for free, we get too many of them - 272 the last time I counted. We do want a subset of them, probably about 5% - 10%.

Desired outcome

• additional examples for metrics
• links to documentation on the data model so I can work out how to write the sampling override