Monitoring

Cloud Watch Metrics

• Provide metrics including CPU Utilization, networking

• Having dimension (instance id, environment)

Cloud Watch Logs

• Containing different log groups representing different applications

• Different log stream including the logs of the instance

• The log can be exported to s3, open search, etc

• Can provide the query to get the desired fields, calculating the result to obtain the insight

• Subscription filter can be set to filter which log can be delivered to destination

• The log agent can be set to EC2 or on premise server to push the log into the stream

Cloud Watch Alarms

• Alarm is used to trigger notification for any metric

• The alarm target can be SNS, triggering auto scaling action

• Composite alarm can be set to set up multiple metric with and / or operator

Cloud Watch Insight

• Container Insight: collect and summarize the metrics and logs from container service (ECS ,EKS, ... )

• Lambda Insight: The detailed metrics of lambda

• Contributors Insight: The "Top-N" contributors through the cloud log

• Application Insight: The insight of all of application including EC2, ECS, lambda, etc

CloudTrail

• Get the history of event/API call made in AWS account

• Can export the logs to S3

• Enable the insight to detect the unusual activities including inaccurate resource provisioning, hitting AWS service limit

• Events are stored for 90 days

AWS Config

• Define the rules for compliance of the AWS resources

• Remediation can be automatically triggered for non-compliant resources, e.g: create jira task...

• Can make good use of event bridge to send notification when there are non-compliant resources