sqlstreams.otel.source.read_success [SQL0102]
Edit this pageAs it appears in your metrics:
sqlstreams.otel.source.read_success
The adapter records whether this collection read retained measurements successfully. Empty results count as success. This gauge is computed by the adapter and is never written to the metrics stream.
Prometheus exposes sqlstreams_otel_source_read_success. A scrape can return
HTTP 200 with this value at 0; check the gauge as well as HTTP availability.
On a read failure, no retained measurement values or rejection count accompany it.
The gauge requires the reader to call SQLStreams’s producer. An already-canceled ManualReader call stops before that point and returns no health gauge. A read that reaches SQLStreams but exceeds its CollectTimeout returns 0 with the error.
For example, installation billing loses its database connection: a scrape
can still succeed, but the gauge reports 0. After the next successful read it
reports 1. A 1 says nothing about the age of the retained observations; use
collector progress for collection liveness.
Looking for a different code? Search every thread.