Kubernetes Performance Monitoring using Amazon CloudWatch Container Insights for AWS EKS.
HTML-код
- Опубликовано: 16 июн 2022
- ===================================================================
1. SUBSCRIBE FOR MORE LEARNING :
/ @cloudquicklabs
===================================================================
2. CLOUD QUICK LABS - CHANNEL MEMBERSHIP FOR MORE BENEFITS :
/ @cloudquicklabs
===================================================================
3. BUY ME A COFFEE AS A TOKEN OF APPRECIATION :
www.buymeacoffee.com/cloudqui...
===================================================================
In this video you will learn and leverage the new CloudWatch Container Insights to see how you can use native CloudWatch features to monitor your EKS Cluster performance.
CloudWatch Container Insights is used to collect, aggregate, and summarize metrics and logs from your containerized applications and microservices. Container Insights is available for Amazon Elastic Container Service, Amazon Elastic Kubernetes Service, and Kubernetes platforms on Amazon EC2. The metrics include utilization for resources such as CPU, memory, disk, and network. Container Insights also provides diagnostic information, such as container restart failures, to help you isolate issues and resolve them quickly.
This video is dedicated for container SME , developer and Operations team to view and troubleshoot the issue at AWS EKS workloads.
Command and Specification file used in video can be found at : github.com/RekhuGopal/PythonH...
#awseks #conatinerinsights #cloudwatch #Observaility #kubernetes 
Наука
thank you so much for the complete guidance.
Thank you for watching my videos.
Glad that it helped you.
It should work. Do follow the same steps again try multiple times (as you might have missed one step)
yes its working i missed attaching the iam role to worker nodes, thank you@@cloudquicklabs
👌any idea on how one can get metric on spot instance termination from the cluster?
Thank you for watching my videos.
When spot instance is also considered as the normal EC2 Instances so, it should as same I explained in my onther video ruclips.net/video/umc752URuIQ/видео.html did you check that
When i check the daemonsets ( kubectl -n amazon-cloudwatch get daemonsets, I have 2 daemonsets created, but with desired = 0, current =0, ready =0, up-to-date=0, Available =0. What is wrong??? please help
Thank you for watching my videos.
This could be issues with the resources of node, can you please check if nodes has enough cpu, memory and disk space for daemonset creation.
Thank you for the guidance. Followed this steps for the performance monitoring. One help. Our application is running in EKS Cluster. Actually, i have created new namespace and deployed application in that namespace. But, this created namespace is not shown in the performance monitoring. Only default, Cloudwatch, ingress, proxy related eks namespaces is visible. How can we include other namespace if application is deploy in eks with different namespace. Please suggest
Thank you for watching my videos.
Indeed you can create workspace.
Did you check the reference file shared in videos description.
@@cloudquicklabs Referred the given details. Created separate workspace for my application. But this workspace (namespace) is not shown in the EKS cluster CloudWatch insights. Do we need to specify our own workspaces in any location of the EnableContainerInsights? Kindy advice.
Fluentd is depricated right. Present we are using fluent bit I think
Thank you for watching my videos.
Fluentd is opensource utility, its not yet deprecated but still new versions are released from community and big companies are still using it.
www.fluentd.org/blog/fluentd-v1.14.6-has-been-released
@@cloudquicklabs thanks for your response. One more question using fluentd will we get container logs, like same as we do kubectl get logs for container.
Thank you for coming back on this.
Yes, all logs are being forwarded to CloudWatch log groups, you can also refer one more video on same at ruclips.net/video/DPCS8jBqmQk/видео.html
Hi My Log group with suffix of /application is not created can you please tell me why, even i added the cloudwatch agent policy on my iam role of worker nodes
Thank you for watching my videos.
Did you follow the configuration as suggested in command file that shared in my GitHub repo : github.com/RekhuGopal/PythonHacks/blob/main/AWS_EKS_Conatiner_Insights/Kubernetes_Container_Insights.json
@@cloudquicklabs Yes... I created my EKS cluster using terraform so I manually added the cloud watch agent on my worked node role, and my region is ap-south-1, so i dont need to change that in EnableContainerInsights.yml I only changed the eks clsuter name and applied the yml but still it is not working and EKS version is 1.24
I am troubleshooting it, and I found this in fluentd logs
2023-06-14 04:14:11 +0000 [warn]: #0 [in_tail_fluentd_logs] pattern not matched: "2023-06-14T04:14:06.70622222Z stdout P 2023-06-14 04:14:06 +0000 [warn]: #0 [in_tail_fluentd_logs] pattern not matched: \"2023-06-14T04:14:04.477041737Z stdout P 2023-06-14 04:14:04 +0000 [warn]: #0 [in_tail_fluentd_logs] pattern not matched: \\\"2023-06-14T04:14:01.976952828Z stdout P 2023-06-14 04:14:01 +0000 [warn]: #0 [in_tail_fluentd_logs] pattern not matched: \\\\\\\"2023-06-14T04:14:01.353621862Z stdout P 2023-06-14 04:14:01 +0000 [warn]: #0 [in_tail_fluentd_logs] pattern not matched: \\\\\\\\\\\\\\\"2023-06-14T04:14:00.241664515Z stdout F
No matter how many I try, I get [error: deployment "understood-zebu-wordpress" exceeded its progress deadline"]. I followed the video as is but still the deployment is not getting rolled out.
Thank you for watching my video.
It could due to insufficient permissions, please check your permissions.
And also did you follow the repo reference document added in description of video github.com/RekhuGopal/PythonHacks/tree/main/AWS_EKS_Conatiner_Insights
How to get the CPU utilisation of cluster using AWS cli ?
Thank you for watching my videos.
You can export the CloudWatch logs to csv directly with using native capability of CloudWatch. And if you would like to have it extracted through CLI you can do so by referring cli commands docs.aws.amazon.com/cli/latest/reference/cloudwatch/index.html
@@cloudquicklabs can you help me out bit more here , I want the node CPU utilisation from last from hour and how to get this metric using AWSCLI .
Thank you for watching my videos.
May request you to please look into my channel membership, I shall provide the required support from there
@@cloudquicklabs Please provide the solution here, Few automations task pending from that.
Thank you again for coming back on this.
Please try using command 'aws cloudwatch get-metric-statistics --namespace AWS/EC2 --metric-name CPUUtilization \
--dimensions Name=InstanceId,Value= --statistics Maximum \
--start-time --end-time --period 360'
Please check link : docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/US_SingleMetricPerInstance.html