cnosdb VS pq

Hi folks! I’m working on a tool for ad-hoc metric extraction and I’m really curious to hear back from the fellow DevOps people. I often find myself in situations when some metrics aren’t collected. Some servers may be simply not onboarded to the metric collection system yet. Some breakdowns may not exist because of the high-cardinality issue, especially per-pod breakdowns. And some metrics we just may haven’t thought about in advance. So, I hacked a tool called pq to parse and query logs as metrics. It seems to work and I have some ideas for further development. For instance, by adding a tiny HTTP adapter to it, it would be able to mimic Prometheus API, so one would be able to attach Grafana to an exposed port and query metrics from the UI. But before going any further with it, I’d like to hear back from the potential users.

Hi fellow SREs! What do you usually do when you need to make an ad-hoc analysis of some metrics that aren't collected yet? For instance, if there is no per-pod request latency measurement, but you want to see how a particular pod behaves? I often find myself in situations like that but I'm not aware of a common solution, so I hacked a tool called pq. It can parse log files into structured streams and then query these steams with PromQL-like language.