dynamic-dynamodb-lambda's Introduction

Dynamic DynamoDB - lambda

Use DynamoDB autoscaling with lambda!
It's Dynamic DynamoDB's simple lambda version

functional/running test passed. but before use, Test it please

please give me any report/suggestion, Welcome

Feature

autoscale DynamoDB's provisioned read/write capacity
run by lambda functon with scheduled event

File Structure

index.js - main handler & flow source. using async
tasks.js - Detail work sources. using AWS SDK
config.js - capacity scaling rule configuration

How to use

$ git clone https://github.com/rockeee/dynamic-dynamodb-lambda.git or download zip
$ npm install to download npm modules
modify config.js for your configuration.

almost same with Dynamic DynamoDB's option

module.exports = {
 region : 'us-west-2', // region
 timeframeMin : 5, // evaluation timeframe (minute)
 tables :
     [
         {
         tableName : 'testTable',     // table name
         reads_upper_threshold : 90,  // read incrase threshold (%)
         reads_lower_threshold : 30,  // read decrase threshold (%)
         increase_reads_with : 90,    // read incrase amount (%)
         decrease_reads_with : 30,    // read decrase amount (%)
         base_reads : 5,              // minimum read Capacity
         writes_upper_threshold : 90, // write incrase amount (%)
         writes_lower_threshold : 40, // write decrase amount (%)
         increase_writes_with : 90,   // write incrase amount (%)
         decrease_writes_with : 30,   // write incrase amount (%)
         base_writes : 5              // minimum write Capacity
         }
         ,
         {
         tableName : 'testTable2',
         reads_upper_threshold : 90,
         reads_lower_threshold : 30,
         increase_reads_with : 0,     // to don't scale up reads
         decrease_reads_with : 0,     // to don't scale down reads
         base_reads : 3,
         writes_upper_threshold : 90,
         writes_lower_threshold : 40,
         increase_writes_with : 0,    // to don't scale up writes
         decrease_writes_with : 0,    // to don't scale down writes
         base_writes : 3
         }
         // additional table...
     ]
};

deploy to lamda function with your favorite method (just zip, or use tool like node-lambda)
check lambda function's configuration

memory - 128MB, timeout - 10sec
set Cloudwatch Event Rule to run your lambda function. for detail, refer this
set & attach role to lambda function
example role policy

 {
 "Version": "2012-10-17",
 "Statement": [
     {
         "Sid": "",
         "Action": [
             "dynamodb:DescribeTable",
             "dynamodb:UpdateTable",
             "CloudWatch:getMetricStatistics"
         ],
         "Effect": "Allow",
         "Resource": "*"
     },
     {
         "Sid": "",
         "Resource": "*",
         "Action": [
             "logs:CreateLogGroup",
             "logs:CreateLogStream",
             "logs:PutLogEvents"
         ],
         "Effect": "Allow"
     }
 ]
}

Roadmap

more stable
add SNS noti when scaled/failed

dynamic-dynamodb-lambda's People

Contributors

Stargazers

Watchers

dynamic-dynamodb-lambda's Issues

Need to Throttle Decreases

You should try to implement something like I have done at https://github.com/bradmurray/DynamoDBLambdaMonitor when it comes to decreases. The problem I ran into in using mine was that all four decreases would usually get used up in the first hour of the day and my savings on provisioning would be minimal. My basic logic is such that you can only use up to 1 decrease in the first 6 hours, 2 in the first 12 and so on. My implementation does not have multi-table as yours does.

Recommend Projects