I have a redis pod with container limits of 6GB. Metricbeat (7.6.2) data from the container shows:
kubernetes.container.memory.usage.bytes: 5.9GB kubernetes.container.memory.usage.limit.pct: 99.9% kubernetes.container.memory.available.bytes: 1.6GB
This container looks to be using 100% of available memory but also has 1.6GB available? We base our alerts on
kubernetes.container.memory.usage.limit.pct so this is problematic right now. Any idea whats going on? I can provide a screenshot if it helps. From some limited sampling, other containers in the clusters correctly show the kubernetes.container.memory.usage.limit.pct as a reflection of the k8s manifest limit.