chrispruitt / aws-switch

An attempt to halt any aws services defined by a given stack to save money and then restart them from their previous state

aws-switch

Use tags to identify aws resources to halt or resume expensive services.

The basic idea:

1. identify resources based on a given set of tags
2. halt - scale resources down to zero and store previous values in a state
3. resume - scale resources back to previous value based on stored state

Currenlty supports

• ecs:services (does not include underlying compute if running on ec2)
• rds:clusters (provissioned only as serverless auto scales on its own)
• rds:db (all database instances that are not a part of a cluster)

Homebrew

brew install chrispruitt/tap/aws-switch

Install from source

Install binary from source.

Usage

    $> aws-switch --help

    aws-switch is used to halt and resume all aws services by tag.

    Usage:
    aws-switch [command]

    Available Commands:
    configure   Creates an s3 bucket for the aws-switch state to reside
    halt        halt an aws service
    help        Help about any command
    resume      Resume a halted aws service

    Flags:
    -h, --help      help for aws-switch
        --version   version for aws-switch

Roadmap

• --exclude-tag flag to optionally exclude resources with specific tags
• different backends - local, dynamodb
• read configuration from yaml file or env vars
• show changes and ask for confirmation before scaling up or down
• flag to select specific services - rds, ecs - default to all
• show state
• resume specific items in state - work around using a good tagging stradegty and just identify resources by tags