Monitor AWS DynamoDB with AppDynamics

[Abhi.Bajaj]

Pre-requisite

1. Machine Agent installed on any Linux box that has access to the DynamoDb Database
2. Linux box should have permission to fetch CloudWatch metrics

Installation

1. If your Linux Box is on EC2, then please select the IAM Role associated with that EC2 instance and add “CloudWatchFullAccess” permission.
   
2. Once done, SSH inside the box where your Machine Agent is running. 
3. In the Machine Agent Home folder, go to the Monitors folder and create a directory called "DynamoDb". In my case, MA_HOME = /opt/appdynamics/ma
   

   cd /opt/appdynamics/ma/monitors
   mkdir DynamoDb​

4. Inside the 'DynamoDb' folder, create a file called script.sh with the below content:
   NOTE: Please edit TABLE_NAME and AWS_REGION with your desired TABLE_NAME and AWS_REGION
   

   #!/bin/bash
   
   # Define the DynamoDB table name
   TABLE_NAME="aws-lambda-standalone-dynamodb"
   
   # Define your AWS region
   AWS_REGION="us-west-2" # Change this to your region
   
   # List of all metrics you want to fetch
   declare -a METRICS=("ConsumedReadCapacityUnits" "ConsumedWriteCapacityUnits" "ProvisionedReadCapacityUnits" "ProvisionedWriteCapacityUnits" "ReadThrottleEvents" "WriteThrottleEvents" "UserErrors" "SystemErrors" "ConditionalCheckFailedRequests" "SuccessfulRequestLatency" "ReturnedItemCount" "ReturnedBytes" "ReturnedRecordsCount")
   
   # Define the time period (in ISO8601 format)
   START_TIME=$(date --date='60 minutes ago' --utc +%Y-%m-%dT%H:%M:%SZ)
   END_TIME=$(date --utc +%Y-%m-%dT%H:%M:%SZ)
   
   # Loop through each metric and fetch the data
   for METRIC_NAME in "${METRICS[@]}"
   do
     # Fetch the metric data using AWS CLI
     METRIC_VALUE=$(aws cloudwatch get-metric-statistics --namespace AWS/DynamoDB \
       --metric-name $METRIC_NAME \
       --dimensions Name=TableName,Value=$TABLE_NAME \
       --start-time "$START_TIME" \
       --end-time "$END_TIME" \
       --period 3600 \
       --statistics Average \
       --query 'Datapoints[0].Average' \
       --output text \
       --region $AWS_REGION)
   
     # Check if metric value is 'None' or empty
     if [ "$METRIC_VALUE" == "None" ] || [ -z "$METRIC_VALUE" ]; then
       METRIC_VALUE="0"
     else
       # Round the metric value to the nearest whole number
       METRIC_VALUE=$(printf "%.0f" "$METRIC_VALUE")
     fi
   
     # Echo the metric in the specified format
     echo "name=Custom Metrics|DynamoDB|$TABLE_NAME|$METRIC_NAME,value=$METRIC_VALUE"
   done​

   If you have multiple tables, then use the script below:
   

   #!/bin/bash
   
   # List of DynamoDB table names
   declare -a TABLE_NAMES=("Table1" "Table2" "Table3")  # Add your table names here
   
   # Define your AWS region
   AWS_REGION="us-west-2"  # Change this to your region
   
   # List of all metrics you want to fetch
   declare -a METRICS=("ConsumedReadCapacityUnits" "ConsumedWriteCapacityUnits" "ProvisionedReadCapacityUnits" "ProvisionedWriteCapacityUnits" "ReadThrottleEvents" "WriteThrottleEvents" "UserErrors" "SystemErrors" "ConditionalCheckFailedRequests" "SuccessfulRequestLatency" "ReturnedItemCount" "ReturnedBytes" "ReturnedRecordsCount")
   
   # Define the time period (in ISO8601 format)
   START_TIME=$(date --date='60 minutes ago' --utc +%Y-%m-%dT%H:%M:%SZ)
   END_TIME=$(date --utc +%Y-%m-%dT%H:%M:%SZ)
   
   # Loop through each table
   for TABLE_NAME in "${TABLE_NAMES[@]}"
   do
     # Loop through each metric and fetch the data for the current table
     for METRIC_NAME in "${METRICS[@]}"
     do
       # Fetch the metric data using AWS CLI
       METRIC_VALUE=$(aws cloudwatch get-metric-statistics --namespace AWS/DynamoDB \
         --metric-name $METRIC_NAME \
         --dimensions Name=TableName,Value=$TABLE_NAME \
         --start-time "$START_TIME" \
         --end-time "$END_TIME" \
         --period 3600 \
         --statistics Average \
         --query 'Datapoints[0].Average' \
         --output text \
         --region $AWS_REGION)
   
       # Check if metric value is 'None' or empty
       if [ "$METRIC_VALUE" == "None" ] || [ -z "$METRIC_VALUE" ]; then
         METRIC_VALUE="0"
       else
         # Round the metric value to the nearest whole number
         METRIC_VALUE=$(printf "%.0f" "$METRIC_VALUE")
       fi
   
       # Echo the metric in the specified format
       echo "name=Custom Metrics|DynamoDB|$TABLE_NAME|$METRIC_NAME,value=$METRIC_VALUE"
     done
   done

5. Great. Create another file called monitor.xml with the below content:
   

   <monitor>
       <name>DynamoDb monitoring</name>
       <type>managed</type>
       <description>DynamoDb monitoring</description>
       <monitor-configuration>
       </monitor-configuration>
       <monitor-run-task>
           <execution-style>periodic</execution-style>
           <name>Run</name>
           <type>executable</type>
           <task-arguments>
           </task-arguments>
           <executable-task>
               <type>file</type>
               <file>script.sh</file>
           </executable-task>
       </monitor-run-task>
   </monitor>​

   
   
   
6. Great work!! Now, Llet’s restart your Machine Agent
7. Once you are done, you will be able to see your DynamoDB metrics in the AppDynamics Machine Agent’s metric browser