Prometheus-Sink - Tags missing

[Schmaetz]

Describe the bug
When using Prom as a sink, tags (labels in Prometheus) seem to be missing from the pgWatch output. This does not happen with all metrics.

To Reproduce
Steps to reproduce the behavior:

1. Add Tag to Target in pgWatch
2. Configure pgWatch for prometheus Sink
3. Check Prom-Sink-Endpoint for
4. See error

curl http://.....:9188/metrics | grep instance_up
pgwatch_instance_up{dbname="instance1"} 1 1747052580658
pgwatch_instance_up{dbname="instance12"} 1 1747052580658

pgwatch_wal_timeline{cluster_name="cluster1",dbname="instance1",sys_id="xyz"} 20 1747052597056
pgwatch_wal_timeline{cluster_name="cluster1",dbname="instance2",sys_id="xyz"} 20 1747052597033

Expected behavior
I expect that the tags are correctly stored in all metrics, not just in a few.

**Version info:
Version:
DB Schema: 00179
Git Commit: f0949f9
Built: 2025-03-25T10:49:31+01:00

Additional context
Add any other context about the problem here.

[pashagolub]

Seems like this is intended behavior for instance_up from previous version, so I remove a bug tag. But I will investigate this issue further on how to improve things

[ShkodaAlexey]

hi, it looks like pgwatch_instance_up is hardcoded somewhere, because in Prometheus it is displayed even when it is not defined in the query, but if you define a metric query "https://github.com/cybertec-postgresql/pgwatch/blob/master/internal/metrics/metrics.yaml#L846" . Then it starts showing them 2 times and for the one that was defined there are custom tags, also if you turn off pg which is being monitored, then the metric that we defined will work correctly.

pgwatch_instance_up{dbname="pgwatch-testdb"} 1 1749023692971
pgwatch_instance_up{dbname="pgwatch-testdb1"} 1 1749023692971
pgwatch_instance_up{cluster="pgwatch",database="testdb",dbname="pgwatch-testdb",env="tst"} 1 1749023687268
pgwatch_instance_up{cluster="pgwatch",database="testdb1",dbname="pgwatch-testdb1",env="tst"} 1 1749023687441