AWS RDS Enhanced Monitoring

See our blog for details.

Plotting

Prototype to plot system metric data collected by AWS RDS Enhanced Monitoring

Dependencies

• aws-cli
• bash
• jshon
• sed
• flot (included in code)

Usage

• define credentials and profile for aws cli

aws configure

• get log group from AWS CloudWatch Logs via AWS Console or cli:

aws logs describe-log-groups --query='logGroups[].logGroupName' --output=text

• get log stream from AWS CloudWatch Logs via AWS Console or cli:

aws logs describe-log-streams --log-group-name <log-group-name>  --query='logStreams[].logStreamName' --output=text

• fetch logs from AWS CloudWatch Logs
  • get all-time data using default profile

bash fetch-logs.sh -g <log-group-name> -s <log-stream-name>

• get all-time data using foobar profile

bash fetch-logs.sh -p foobar -g <log-group-name> -s <log-stream-name>

• get data from 1454573815 (Thu Feb 4 09:16:55 CET 2016) until 1454574007 (Thu Feb 4 09:20:07 CET 2016) using foobar profile

bash fetch-logs.sh -p foobar -g <log-group-name> -s <log-stream-name> -b 1454573815 -e 1454574007

• get data from 1454573815 (Thu Feb 4 09:16:55 CET 2016) until now using foobar profile

bash fetch-logs.sh -p foobar -g <log-group-name> -s <log-stream-name> -b 1454573815

• open index.html in your browser

Adding graphs

• add function call to get wanted metric to fetch-logs.sh
• add plotting segments to javascript of index.html
• for available metrics see AWS RDS documentation
• included metrics:
• diskIO
• avgQueueLen
• await
• readIOsPS
• readKbPS
• tps
• writeIOsPS
• writeKbPS
• util
• cpuUtilization
• wait
• irq
• system
• steal
• loadAverageMinute
• five
• memory
• buffers
• cached
• dirty
• writeback
• network
• tx
• rx

TODO

• implement array awareness for diskIO and network