I’m trying to introduce an overall rate limit on our span ingestion, to avoid high error rates (which we normally want to sample at 1.0) elsewhere taking down the monitoring system as well. spans_per_second seems to be exactly what I want, but for the tail_sampling processor it seems that spans_per_second needs to be part of every existing policy, of which 1) there are a lot and 2) I’d like an overall limit applied after the other policies are evaluated.
Is there some clever way to do this after other policies are applied? Or is there a similar option in another part of the stack that I could use instead?