The new infra monitoring section reports much higher memory usage compared to kubectl top command. what might be the reason, is it a bug? e.g. max memory usage for nodes when checked from kubectl = 52%, and on signoz = 85%

so just to be clear, you guys are using different method from metric server for signoz dashboard? what I am not understanding is how there is so much difference b/w metrics if signoz and metric server are using similar approach

usage by metric service heuristic and doesn't include all cache memory; usage shown in infra monitoring comes from the underlying operating system