I have some alerts set up via Prometheus for latency based on the qos_latency stat. I noticed when I went onto the Harvest Volume dashboard however that it didn't show up in the top 5 volumes by QoS latency when I set the dashboard to look at the last 30 minutes. Reviewing the dashboard source it looks like the top N is based on the average over the last 3 hours always, not on the last e.g. 15 minutes. Is there a reason for that? Or am I missing something?
#Dashboards Top N hard coded to last 3 hours
pseudo ember
heady vigil
@pseudo ember Yes, top-k in dashboards currently uses a hard-coded 3-hour window. While we could derive it from Grafana’s selected time range that approach doesn't scale well for long durations like multi-month selections. Queries are very slow or time out in such cases.