The coverage in MetricsHub shown in Prometheus is not accurate

[MedMaalej]

In Prometheus Dashborads, we display a percentage of coverage, which means the hosts that are really monitored compared to the hosts that are configured (in the YAML file metricshub.yaml).

The percentage of coverage in MetricsHub is very variable compared to HWS, the metric metricshub_connector_status value is not accurate.

The coverage in Metricshub changes from 18% to more than 90% at once. In HWS, it's mostly between 95% and 97 %.

Some hosts are sometimes monitored and sometimes not monitored.

• metricshub_connector_status in Metricshub:

• hardware_sentry_connector_status in HWS:

Here is an example of the problem occurring with the host anaxagore (windows):

• In MetricsHub

• In HWS

We need to find the root cause and fix the problem