This project provide a sytem structure to stop AWS RDS resources continuously.
Currently(As of 2019-04-25), you can stop DB instances / clusters that are RDS resources, but they will be started automatically after 7 days.
In order to stop the DB instance / cluster continuously under this restriction, it is realized using the following AWS services.
Lambda is used to execute a function to stop a resource that has been tagged to indicate the stop target. And CloudWatch Events is used to invoke this function daily at a.m. 0 (default) in UTC time zone.
The above configuration is built using Serverless Framework.
rds_stopper
├── README.md
├── lambda … Lambda function scripts.
│ └── main.py
└── serverless … Serverless Framework settings.
├── events … CloudWatch Events settings.
│ └── main.yml
└── serverless.yml
3 directories, 4 files
verificated runtime envirnment as following.
$ sls --version
1.36.3
- Python version: 3.7.2 (default, Mar 1 2019, 11:28:42)
- boto3 version: 1.9.42
First time, you have to execute deploy command.
$ cd /path/to/rds_stopper/serverless
$ sls deploy
If you ware changed only lambda function, you will be able to deploy function command.
$ sls deploy function -f main
Using AWS console / AWS CLI, Add the following tag(Key/Value) to the resource which you want to stop.
- Key
- LAMBDA_RDS_STOPPER_TARGET (default)
- Value
- True (A string representing 'true', for example True/true/1...)
If you would like to change Key name, you can change TARGET_TAG in serverless.yml.
$ cat serverless.yml
…
TARGET_TAG: LAMBDA_RDS_STOPPER_TARGET
…
If the RDS resource is a cluster, you have to tag the cluster. Even if you tag individual instances belonging to a cluster, you can not stop them alone.
basic invokation is as following.
$ sls invoke -f main
If you would like to change the log level, you can set "log_level" in Lambda event data.
$ sls invoke -f main -d '{ "log_level": "DEBUG" }'
valid "log_level" string are as following.
- DEBUG
- INFO
- WARNING
- ERROR
- CRITICAL