Related to TRT-849, we want to write a test to see how often this is happening before we undertake a major effort to get to the bottom of it.

The test will need to process disruption across all backends, look for DNS lookup disruptions, and then see if we have overlap with non-DNS lookup disruptions within those timeframes.

We have some precedent for similar code in KubePodNotReady alerts that we handle differently if in proximity to other intervals.

The test should flake, we can then see how often it's happening in sippy and on what platforms. With sql we could likely pinpoint to certain build clusters as well.