Management Tools > CloudWatch

CloudWatch is available for several different AWS services

• EC2
• ELB
• S3
• DynamoDB
• RDS
• Etc.

• Basic Monitoring every 5 minutes
• Advanced (Detailed) Monitoring every minute – but is not free.

Dashboards

Dashboard widgets, outside of Text, allow you to add as many ‘metrics’ to a graph (Line or Stacked) or Number as you wish.

• Dashboard > [Create Dashboard]
• Widgets (4)
  • Text (Markdown text)
    • [Add Widget] > Text > [Configure]
    • Uses Markdown language
      • # Heading 1, ## Heading 2
      • [link text](http://url.com)
      • *italics*, **bold**, ~~strikethrough~~
  • Line (Basic Line)
    • [Add Widget] > Line > [Configure] > Select service type (EC2) > ‘Per-Instance Metrics’
      • Available metrics listed by Instance Name / ID
    • Select Metric *Know these for the exam!
      • CPU Related
        • Credit Balance
        • Credit Useage
        • Utilization
      • Disk Related
        • Read Bytes
        • Write Bytes
        • Read Ops
        • Write Ops
      • Network Related
        • In
        • Out
        • Packets In
        • Packets Out
      • Status Check
        • For the Instance
        • For the Hypervisor (Host)
      • If a Metric is not listed, you will need to create a “Custom Metric” for it.
        • RAM Utilization
        • Disk Space Available
    • Stacked Area (Same as Line, but lower half colored in)
    • Number (actual value of Metric during last poll)

Alarms

Unlike Dashboard widgets, Alarms only allow you to set trigger points on a single specific Metric.

• Select Metric
  • Select the metric you wish to set the trigger point for > [Next]
• Create Alarm
  • Alarm Threshold
    • Name (Must be unique?)
    • Description -> This is the alert that will be sent
    • Set Trigger value (<, <=, >=, >)
    • Set ‘Consecutive Periods’ (A period is the time delays between polls.  Standard is 5 minutes.
  • Additional Settings
    • Treat Missing data as:
      • good (not breaching threshold – will clear an active alert)
      • bad (breaching threshold – will cause an alert)
      • ignore (maintain current alert status)
      • missing (default – definition not clear)
  • Actions
    • You can add as many ‘Actions’ as you wish
    • Notifications
      • Whenever this alarm: (Select Alarm state)
        • State is ALARM
        • State is OK
        • State is INSUFFICIENT
      • Send notification to: (Notification List)
        • All Recipients must ‘Opt In’
        • New List
          • Enter Topic Name (seems like a ‘Key’ type value.  No spaces!)
          • Add Email Recipient(s?)
            • separate with semicolon?
    • AutoScaling Action
    • EC2 Action
      • Whenever this alarm: (Select Alarm state, see above)
      • Take this action
        • Recover this instance (Reboot?)
        • Stop this instance
        • Terminate this instance
        • Reboot this instance
      • * AWS will automatically create a new IAM Role to allow these actions!
• Events (Understand at high level)
  • Respond to state changes to AWS resources.
  • Examples
    • Automatically invoke an AWS Lambda function to update DNS entries when an event notifies you that Amazon EC2 instance enters the Running state
    • Direct specific API records from CloudTrail to a Kinesis stream for detailed analysis of potential security or availability risks
    • Take a snapshot of an Amazon EBS volume on a schedule
• Logs (Understand at high level)
  • Examples
    • Monitor HTTP response codes in Apache logs
    • Receive alarms for errors in kernel logs
    • Count exceptions in application logs
  • Install the Agent to send your logs to the CloudWatch Logs service.
  • Create metric filters to automatically monitor the logs that are sent.
  • View the log data you sent and is stored in CloudWatch Logs.
• Metrics
  • Easy.  Use this instead of creating dashboards.

Exam Tips

• Standard Monitoring every 5 minutes
• Detailed Monitoring every 1 minute
• What can you do?
  • Create dashboards
  • Set Alarms
  • Trigger Events based on rules
  • Aggregate, monitor and store logs.
• Difference between CloudWatch and CloudTrail
  • CloudWatch is for MONITORING
  • CloudTrail is for AUDITING (Who did what, when and where)