I have 20 hosts and need to configure alert “cpu > 90%”.
If I configure this on graf I don’t get node name on which CPU has load > 90%.
If I create separated graf-and-alert for every host - it works fine for me, but
I don’t want to add new graf-and-alert for every new node.
Are there some solution which has both benefits?
What data source are you using? The notification should include the series name that triggered the threshold / condition evaluation.
it contains cpu.percent metric name and value, but not hostname.
when the OpenTSDB alerting support was contributed they missed proper alias / series naming,
Sadly the main community OpenTSDB contributor is no longer working on Grafana stuff so this bug has been left open for a while.
I am also Kind of blocked by the same issue (DataSource is ELASTICSEARCH 6.7 version). Although my use case is different, but its the same issue that I don’t get which executor’s heap memory crosses the threshold.