How to create a CloudWatch Alarm for CPU utilization?

[@mathewkargarzadeh3158]

you folks are amazing !!. I really really appreciate for your time sharing your knowledge has so much humanity in it. teaching every one free. your are very much appreciated on my part !!

[@ccm590]

this is not Pytholic, it is alcoholic...I am addicted bro

[@ccm590]

a BIG heart for you bro

[@vigilkumarap]

@pythoholic: Thanks to ur channel, was able to get my basics clear and clear the exam yesterday.

[@Pythoholic]

Awesome news,,, congratulations 👍 please help the channel if u wish to by posting ur accomplishment on linked in and tagging the channel. Many congratulations

[@honoratob.felicitasjr.7529]

Thank you for this video. I was able to create a CPU alarm for Redshift for the first time. :)

[@matrixtoogood5601]

Amazing explanation, your channel will become very popular for all AWS certifications in a few years!

[@pannalal3994]

pls share these wonderful piece of knowledge n effort which will keep motivated the teacher...

[@amritasingh1105]

Nice tutorial with clear explanation 👌

[@SravyaTirumala]

Thankyou for the video. It was so clear and have understood clearly.

[@TechJedi007]

Not sure why, but I created the alarm, created a topic, and subscribed to the topic. I'm still unable to get any notifications even though the metrics I set were really low after running the stress test.

[@ShadabAnsari-rc7cz]

very good and informative lecture. u r rocking man.

[@joeldjomou4071]

Awesome video man!!! Thank you for the good work.

[@kaboodlefish]

Is there a way to prevent ec2s from contributing to metrics until x minutes after they've started? I'm deploying via terraform, so my cpu alarm is sounding whenever an ec2 starts as it's obviously using a lot during that initial period.

[@Pythoholic]

Yes, you can prevent newly launched EC2 instances from contributing to metrics until a certain amount of time has passed after their start. One way to achieve this is by using a combination of AWS CloudWatch Alarms and Auto Scaling Lifecycle Hooks. Here's a high-level overview of the process: 1. **CloudWatch Alarm**: Create a CloudWatch alarm to monitor the CPU usage of your EC2 instances. This alarm will be triggered when the CPU usage exceeds a certain threshold. 2. **Auto Scaling Group (ASG) Configuration**: Ensure that your EC2 instances are launched within an Auto Scaling Group (ASG). The ASG is responsible for maintaining the desired number of instances and automatically launching or terminating instances based on your defined policies. 3. **Lifecycle Hook**: Create an Auto Scaling Lifecycle Hook that pauses the EC2 instance when it's launched. A lifecycle hook allows you to perform custom actions as instances launch or terminate within the ASG. 4. **Pause EC2 Instance**: When the EC2 instance is launched, the lifecycle hook will be triggered, and the instance will be in a "wait" state. During this time, you can implement a script or mechanism on the instance to prevent metrics reporting (such as disabling CloudWatch agent or other monitoring tools). 5. **Delayed Metric Contribution**: After a specified amount of time (e.g., x minutes), you can have your custom script or mechanism enable monitoring tools and allow the instance to start contributing to metrics. At this point, the lifecycle hook can complete, and the instance will be fully operational. By using this approach, you can avoid false alarms for high CPU usage during the initial period after launching EC2 instances. Keep in mind that you will need to handle the lifecycle hook logic in your EC2 instance's userdata or launch script. You can use AWS SDKs or CLI commands in your userdata script to interact with Auto Scaling and signal when the instance is ready to start contributing to metrics. Implementing this solution requires some scripting and configuration work, but it can be a useful way to prevent alarms from triggering unnecessarily during the initial period after launching EC2 instances.

[@bibifedomatova2115]

while creating ec2 why did you use one of your own subnet instead of default one. is there any purpose behind this? Your demos and graphic demonstration is on another level. thank you! its educational and fun!

[@Pythoholic]

Yes its because we wanted to host it on our network ie the one we created to test it.

[@leovillaz]

I´m trying to configure Cloudwatch to monitor CPU utilization per thread but I only catch CPU utilization average. I would like to know if it is possible to configure Cloudwatch to monitor each of the CPU threads individually, as MSSQL runs on the server, which is configured to use 6 of the 8 available threads. I looked through various AWS documentation and none, in particular, told me how to configure Cloudwatch for this. And even using the Cloudwatch Agent I couldn't get the expected result.

[@Pythoholic]

Have you tried cloud watch metric "ThreadUtilization" did that help: cloudwatch = boto3.client('cloudwatch') # Define the metric data metric_data = [ { 'MetricName': 'ThreadUtilization', 'Dimensions': [ { 'Name': 'InstanceId', 'Value': 'i-1234567890abcdef0' }, { 'Name': 'ThreadId', 'Value': '1' }, ],

[@diptidahat1141]

How do I create a CloudWatch alarm for CPU Utilization metric for multiple instances? It lets you select just one instance while creating an alarm but I want to add multiple instances.

[@Pythoholic]

You can refer to composite alarms : docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/Create_Composite_Alarm.html But can you please share your requirement, in what condition do u plan to monitor multiple instance for one value

[@vinothkumarak8880]

I have a query, if want create alarm by single metric itself for high cpu utilization more than one instances? if is possible, please help me on this.

[@Pythoholic]

Yes, it is possible to create an alarm based on a single metric for high CPU utilization across more than one instance. This process will largely depend on the system or service that you are using to monitor your instances, such as AWS CloudWatch, Google Cloud Monitoring, or some other monitoring service. Here's a general outline of how to do it in AWS CloudWatch: 1. **Create a new Alarm:** In the CloudWatch console, navigate to Alarms and click on "Create Alarm." 2. **Select a Metric:** Search for and select the `CPUUtilization` metric associated with your instances. If your instances are part of an Auto Scaling group, you can use the `CPUUtilization` metric associated with the Auto Scaling group. 3. **Define the Alarm condition:** You'll need to specify the threshold for the alarm. For high CPU utilization, you might set this at 80%, for example, but the exact figure will depend on your requirements. 4. **Configure Actions:** Define what action should be taken when the alarm state is triggered. This could be sending a notification, triggering an Auto Scaling action, etc. 5. **Add a Name and Description:** Give your alarm a name and a description so you can easily identify it later. 6. **Create the Alarm:** Finally, review your settings and create the alarm. This is a very basic example and the exact steps can vary based on your specific requirements and the monitoring service you are using. Please let me know if you need help with a specific monitoring service or if you have more detailed requirements.

[@saeedshaikh4369]

I have EKS cluster and would like monitor all EC2 instance under that , how to achieve this without applying separate alarm on each instance

[@Pythoholic]

docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/deploy-container-insights-EKS.html This can help you setup insight for EKS cluster.

[@Citypoint888]

Thanks for the awesome video. You are the best.

[@ganeshwankhede3329]

For connecting to EC2, which UI tool you are using ? alternative to Putty ?

[@Pythoholic]

Termius .. i have a video where i have shown how to use it Mobaxterm Putty Termius please check that

[@ganeshwankhede3329]

@@Pythoholic thanks

[@HA04259]

Nice explanation.. when states will change Alaram to OK and OK to Alaram?

[@Pythoholic]

When the condition are met for the trigger of the alarm

[@yadikierukulareddy3280]

Hi , Cant we create a single alarm for metric of cpu utilisation for couple of ec2 instances ?

[@Pythoholic]

Why not? Yes you can

[@m.vineeth9724]

Which software did you use for connecting to the EC2 instance ?

[@Pythoholic]

Termius.. I have a video explaining how to use it as well.

[@m.vineeth9724]

@@Pythoholic Thanks a lot, I ll definitely go through that video.

[@p1xedge]

please upload video how to set alarm in cloudwatch and scale ec2 instance as load increases and send sns notification using python boto3

[@Pythoholic]

Sure

[@brucewayne-bv4ok]

Hey small doubt why did u keep 0.6 as threshold does that mean if cpu utiliZation is above 60 % u will be notified ??

[@Pythoholic]

Yes that was just for a demo purpose you can change it as per your requirement

[@brucewayne-bv4ok]

@@Pythoholic so do I have to keep it 75% if I want to get notified if cpu utilisation crosses 75%??

[@Pythoholic]

yes and based on that number of thresholds you keep , for example it may be like 3 times then once it crosses for 3 times then it will notify

[@brucewayne-bv4ok]

@@Pythoholic hey one more thing !! what about the average , sum , max and min !! do they play a factor ?? am keeping avg jst like u did

[@Pythoholic]

you can read this it would take be easier for understanding docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/Statistics-definitions.html

[@rohithreddykatta8614]

We have around 50 lambda function’s and for them we have to setup “Errors” metric cloud watch alarms. Looking for automation options.

[@Pythoholic]

terraform is the way

[@byronmedina6150]

funciona webillo 10/10

[@samitingole8672]

which software use for cannet ec2 insta

[@Pythoholic]

Putty mobaxterm for windows Ssh for Linux

[@gungunsingh7274]

Thank you sir

[@vinothkumarak8880]

same @dipti dahat asked the query, that could you please clarify me.

[@sharukhshaikh8355]

With what else we can do monitorining

[@Pythoholic]

Monitoring, Alert, Reporting, Dashboards, Event based notification and automation.

[@uploadvoice]

Kibana (ELK stack)

[@manoraja9356]

Please do on events

[@abalilziad4444]

Hello , do you have a udemy courses ?

[@Pythoholic]

Hi I am working on the udemy course will update on the channel once done. Thanks for the feedback

[@abalilziad4444]

@@Pythoholic I m waiting for it , thanks

[@pabitrokumar3686]

Great