Metrics for Prometheus are no longer returned when polling devices

[valerarar]

Metrics for Prometheus are no longer returned when polling devices a junos_exporter
This happens even when polling 10 devices
As a result, the alerts will no longer fire:

Is it possible to store data statically in a file permanently and update as new values come in?

[lethalwp]

I haven't checked this particular issue. But is it possible that your polling rate is too low? So prometheus "forgets" about the metric. If it is this problem, the solution is either to poll more frequently, or change a prometheus option, i think it is: --query.lookback-delta=15m which i've set at 15minutes for my usage case

[valerarar]

No, the problem is that while the metrics are being collected, the current metrics become unavailable. And when the alert manager polls the value, it gets no data.