How to ignore warmup errors?

I have a scenario which I am not able to grapple my hands around. Please help,

So the problem is I have some data coming in at certain time of the day, and when it comes in the machine is warming up and so we get a lot of errors at that time, you can see the spikes in graph below, those spikes are expected and they cause unnecessary alerts, how can I ignore those warmup cases and only raise alerts after that error burst has passed?

I was thinking increasing the time interval for alert generation is a straightforward solution to this, but I think these kind of warm up cases might be common and there might be a better solution out there for solving this.

Hi, @mi-headless: Do these spikes always happen at the same times of day? If so, you might try creating a baseline alert condition. The New Relic algorithm will notice that these spikes occur regularly and not create an alert incident, because they are expected behavior.


Yes these are around the same timeframe, let me explore this topic.

I like @philweber 's suggestion. The only other thing I can think of is a scheduled Muting Rule that silences notifications during that time window?

1 Like