Grafana alert configuration

greetings to the forum …

one of my metrics is figuring a system_runtime - which should constantly increase by time. in case of an error_condition the value is kept constant.

how should be the rule configured to flag an alarm when matrics value is constant (or missing)?

with the given condition-ruleset:
WHEN avg () OF query (A, 5m, now) IS BELOW|IS OUTSIDE RANGE|…
i can’t find a solution …

thank you for helping
linklocal :slight_smile:

meanwile i found a solution …
when everything is working properly the gauge-value: weider_runtime is representing the runtime in [s]. so its constantly rising.

when communication to the device is interrupted or fail - value kept constant (if prometheus-node-exporter still running) or is absent “nan” (if prometheus-node-exporter was killed and unable to restart again)

by differentiating this graph you get a constant value when everything is OK.

which could be evaluated simple by the possible alert-rules
maybe this solution is not perfect but actually it’s good enough (beside some artefacts from rating with unsynchronized timestamps)

… sorry, for asking
linklocal :slight_smile: