Main Menu

Section Menu

• Introduction to monitoring and alerting for AWS RDS
• Monitoring AWS RDS metrics
• AWS RES events
• Quiz

Introduction to monitoring and alerting for AWS RDS

https://www.udemy.com/aws-monitoring-alerting-with-aws-cloudwatch-and-aws-sns/learn/v4/t/lecture/7082820?start=0

Overview

• Intro to RDS
• Key Metric available
• Monitoring for availability and performance

AWS RDS

• Relational Database Service
• Provides a scalable database service
• PaaS – Platform as a service
• Multiple Database engines supported
  • MySQL
  • PostGre
  • MSSQL
  • Oracle
  • MySQL Clone called Aurora
• Highly scalable for large data loads
• Can be configured with High Availability mode
  • Automatically fails over to a healthy replica

Monitoring AWS RDS

• Proper monitoring and alerting is essential
• RDS and CloudWatch provide various monitoring options
  • Provided Metrics
  • Alarms
  • Logs
  • Events

Demo – Overview of RDS Monitoring

• Events
  • Events related to the database instance
    • Includes the time and date of the event
    • Source name and source type
    • Message associated with the event
    • Can be configured to send notifications
  • Database
  • DB Snapshots
  • DB Security Group changes
  • Parameter Groups
• Event Suscriptions allow you to define the type/method of sending alerts.
• Logs:RDS > Instances > Select instance > Scroll way down to Logs
  • Available logs depend on the type of DB engine you are using.
    • MySQL
      • Error Log
      • Slow Query Log
      • General Log
  • Logs are updated every 5 seconds

Monitoring AWS RDS metrics

https://www.udemy.com/aws-monitoring-alerting-with-aws-cloudwatch-and-aws-sns/learn/v4/t/lecture/7082822?start=0

Key Metrics

• CPU Utilization: % of CPU utilized
• *Database Connections: Number of connections currently in use by all the clients of the database.
• Freeable Memory: Ram Available (MB)
• *Free Storage Space: Storage Space Available (MB)
• Read IOPs / Write IOPs
  • High IOPs may suggest volume backing the database may get overloaded
    • This can cause slower responses
• *Read Latency / Write Latency: Track the average time taken per disk input, output or IO operation.
  • High latency values are a cause of concern because it is an indicator that:
    • disk activity is overloaded
    • performance is degraded

* These are the most important when it comes to production monitoring!

Demo: RDS Metrics

• RDS > Instances > Select Database Instance
  • Demo is old and does not match current AWS display 🙁

Create Alarm

• RDS > Instances > Select Database Instance > Scroll down to “CloudWatch Alarms” > [Create Alarm]
• Typical alarm setup
  • Where to send the notification
  • Metric to monitor
  • Thresholds
  • Alarm name

AWS RDS events

https://www.udemy.com/aws-monitoring-alerting-with-aws-cloudwatch-and-aws-sns/learn/v4/t/lecture/7082824?start=0

Key metrics should be monitored closely

• Resource Utilization
• System errors
• Accidental termination
• Cluster health
• Maintenance windows
• Query logs
• RDS metrics
• Instance level metrics
• RDS resource modification

RDS Events

• RDS > Events
  • Generated whenever there is a new configuration change or important operational change from within the RDS instance.
• You can retrieve events from RDS
  • Console: 24 hr
  • RDS API: 14 Days
  • AWS CLI: 14 Days

Creating an Event Subscription

• RDS > Event Subscriptions > [Create Event Subscription]
  • Details:
    • Name
  • Target:
    • Send notifications to: Select ARN (Topic)
  • Source:
    • Source Type:
      • Instances
      • Security groups
      • Parameter groups
      • Snapshots
      • DB clusters
      • DB cluster snapshots
    • Instances to include:
      • All instances
      • Specific (Select from dropdown)
    • Event Categories
      • All
      • Select specific
        • availability
        • backtrack
        • backup
        • configuration change
        • creation
        • deletion
        • failover
        • failure
        • low storage
        • maintenance
        • notification
        • read replica
        • recovery
        • restoration
    • Note: by selecting All Instances and All Categories, we should be notified on any change.
    • [ Create ]
      • This took SEVERAL minutes

Testing

• Reboot the instance
  • It took a while before the notifications came through. Might be a yahoo thing.

Recap:

• Learned various metrics available
• Setup Alarms and Notifications
• Viewed Events and create an Event Subscription
• Viewed logs

Quiz

What does RDS stand for

• Rational Database Service
• Relational Database Service
• Resolution Database Service
• Revolution Database Service

Which of the following key metrics are available for AWS RDS

• CPUUtilization, DatabaseConnections
• FreeMemory, FreeStorageSpace
• ReadIOPS/WriteIOPS
• All of the above

For Monitoring AWS RDS databases, CloudWatch provides:

• Only CloudWatch Alarms
• Only CloudWatch Logs
• Only CloudWatch Events
• All of the above

Which of the following metrics is used to check the number of database connections in use:

• DiskQueueDepth
• DatabaseConnections
• ReadThroughput
• WriteThroughput

The FreeStorageSpace metric is used to

• Determine the amount of currently available storage space
• Determine the amount of available Random Access Memory
• Determine the number of outstanding IOs waiting to access the disk
• Determine the amount of swap space used on the DB instance.