-
Hi, On documentation page I see that this resource will be deprecated. I think that it is really easy to use when it comes to alerting when synthetics monitoring check is failing. The recommendation -> nrql check - is much more difficult to use in this case (I am new to New Relic). Can you tell me why this is deprecated and maybe some idea how can I achieve same result using nqrl check (automatic resolver, fast alerting) |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 3 replies
-
Hi @kirsie I've reached out internally for some examples. |
Beta Was this translation helpful? Give feedback.
-
Hi @kirsie - with NRQL alerts, especially for Synthetics - it's complicated. The problem is the cadence at which synthetics runs. If you have checks running minutely, that's fine. But the moment you switch to less frequent runs, then alerts NRQL gets hugely more difficult. Therefore all examples will end up being complicated. To the point where you need to calculate every element of a NRQL condition. Window durations need to match the frequency of the synthetic check. Which means there will likely be an extended delay between a failure occurring and an alert opening up. But, an example that you could use is, for an alert that triggers when a monitor fails, and the monitors frequency is set to 15 mins:
So, something like the attached screenshot should work. But - this should be set up with the caveat that as with any data signal that's not an easy constant stream of data, the alerts will likely require some iteration to get perfectly right. |
Beta Was this translation helpful? Give feedback.
Hi @kirsie - with NRQL alerts, especially for Synthetics - it's complicated.
The problem is the cadence at which synthetics runs. If you have checks running minutely, that's fine. But the moment you switch to less frequent runs, then alerts NRQL gets hugely more difficult.
Therefore all examples will end up being complicated.
To the point where you need to calculate every element of a NRQL condition. Window durations need to match the frequency of the synthetic check. Which means there will likely be an extended delay between a failure occurring and an alert opening up.
Just by nature of the fact that the data is always choppy. With checks running at a cadence, it's never a smooth simple …