problem with HA prometheus and promscale

[hvulin]

If you are using prometheus with labels cluster and replica as required for data deduplication using promscale with --highavailability this exporter overwrites the system label for cluster and suddenly you have duplicate data in the timescaledb because more than 1 node is labeled as master in ha_lease table... (which then breaks graphs in grafana)... Could you use some othen label from elasticsearch and not cluster since it breaks prometheus.