Advisor2 hours ago, mnagel said:A related change for alerts that would not be solved by inheritance but I had also benefited from in our previous tool is threshold calculation over time.
this is also how one of our old monitoring tools, CA Performance Management (and CA eHealth before that) used to do - calculate over time instead of consecutive polls. that way occassional spikes are ignored but picked up issues if spikes are frequent.
e.g. if utilization value is above threshold for 30min out of 1 hour window (fixed or sliding etc.), then generate alarm.
Dynamic thresholds in LM can somewhat cover for this, i think, but still not as flexible as defining our own 'threshold violation duration' in a 'specified window'
