I’ve been looking at Hastic and LoudML Grafana plugins and their associated servers to leverage ML to identify anomalies in monitoring. On the face of it, as PMM is based on Grafana, they should work, but the interface is different enough that some of the documentation for Hastic at least refers to panels and configs that don’t appear to be present in PMM.
Has anyone tried implementing these tools?
It would be really cool to use anomaly detection to trigger alerts rather than arbitrary thresholds