Hi everyone, I have setup like this: Open Observe is deployed on the AWS EC2(open-source-version, single node), and my application is running on the "AWS ECS" and i am sending the metrics via open Telemetry(running as a sidecar container). To get the CPU utilization and mem utilization, in open observe i am using these queries: Query for CPU Utilization: (ecs_task_cpu_usage_vcpu{aws_ecs_service_name="tr-fantasticwilds-dev-ft-svc"} / 0.25) * 100 {here 0.256 = max CPU that task can use, defined in the task definition} Query for memory utilization (ecs_task_memory_utilized{aws_ecs_service_name="tr-fantasticwilds-dev-ft-svc"} / ecs_task_memory_reserved{aws_ecs_service_name="tr-fantasticwilds-dev-ft-svc"}) * 100 but the problem is that I am getting different CPU utilization graph in CloudWatch and OpenObserve, i have doubt like am i using wrong query to find the cpu utilization in openobserve can anyone please help

Query seems fine, overall trend is similar as well. can you check for other data points as well this will help us conclude if their is a delay or metrics are missing for a specific point.

yes, sometimes data points in openobserve and cloudwatch are same and sometimes they are different by 20-30%. yesterday they were different by 30%.

Parth, I could not think of a very specific reason why this is happening. Although I would like to try one thing: We have a direct integration for cloudwatch - You can find the exact steps for it under Data Sources -> AWS -> Individual Services -> CloudWatch metrics If you can create a new org and ingest these metrics via cloud watch directly, removing the otel layer - this should give us some clarity.