“This distinction follows directly from how Kubernetes enforces resource limits.”
“For example, AWS Compute Optimizer supports 14-, 32-, and 93-day lookback periods, depending on configuration.”
“For example, AWS Compute Optimizer supports 14-, 32-, and 93-day lookback periods, depending on configuration.”
“AWS Compute Optimizer, for example, considers resource specifications such as CPU, memory, EBS throughput, IOPS, disk throughput, and network throughput when calculating performance risk.”
“AWS Compute Optimizer accounts for this issue: when memory metrics are unavailable, it attempts to avoid recommending a reduction in that resource dimension.”
“An interesting real-world example is Amazon EKS Auto Mode resource optimization.”
“CPU P95 usage at 87% of the recommended request. - Memory P95 usage at 70% of the recommended request.”
“AWS EKS Auto Mode provides an interesting example of a feedback mechanism: it increases memory recommendations after OOMKills, with progressively larger increases if further kills occur within 24 hours.”
“For Kubernetes, this can be combined with Vertical Pod Autoscaler (VPA), Horizontal Pod Autoscaler (HPA), and node autoscaling.”
“For Kubernetes, this can be combined with Vertical Pod Autoscaler (VPA), Horizontal Pod Autoscaler (HPA), and node autoscaling.”
“- Prometheus and kube-state-metrics.”
“- Prometheus and kube-state-metrics.”
“CloudWatch doesn't automatically provide comprehensive EC2 memory utilization metrics without additional instrumentation, such as the CloudWatch Agent.”
“CloudWatch doesn't automatically provide comprehensive EC2 memory utilization metrics without additional instrumentation, such as the CloudWatch Agent.”
“CloudWatch doesn't automatically provide comprehensive EC2 memory utilization metrics without additional instrumentation, such as the CloudWatch Agent.”
