Add metric name to metadata error

[gouthamve]

Currently our errors look like this:

level=warn ts=2021-07-14T09:33:02.743961441Z caller=ingester.go:681 org_id=XXXX traceID=3224439dfe4ee37c msg="failed to ingest some metadata" err="per-metric metadata limit of 10 exceeded, please contact administrator to raise it (local limit: 10 global limit: 0 actual local limit: 10)"
level=warn ts=2021-07-14T09:33:02.743862184Z caller=ingester.go:681 org_id=XXXX traceID=3224439dfe4ee37c msg="failed to ingest some metadata" err="per-metric metadata limit of 10 exceeded, please contact administrator to raise it (local limit: 10 global limit: 0 actual local limit: 10)"
level=warn ts=2021-07-14T09:33:02.743728652Z caller=ingester.go:681 org_id=XXXX traceID=3224439dfe4ee37c msg="failed to ingest some metadata" err="per-metric metadata limit of 10 exceeded, please contact administrator to raise it (local limit: 10 global limit: 0 actual local limit: 10)"
level=warn ts=2021-07-14T09:33:02.743496095Z caller=ingester.go:681 org_id=XXXX traceID=3224439dfe4ee37c msg="failed to ingest some metadata" err="per-metric metadata limit of 10 exceeded, please contact administrator to raise it (local limit: 10 global limit: 0 actual local limit: 10)"

Knowing the series name would be helpful.

[pracucci]

LGTM. It's not the most optimized way to do it (for sample ingestion errors we did a large work to optimize the error path) but metadata ingestion errors didn't show to be a bottleneck so far. I'm also suprised there's no unit test error to change, but CI will eventually show it in case.