1.0.4 • Published 2 years ago

@xoi/serverless-log-dumpster v1.0.4

Weekly downloads
-
License
Apache-2.0
Repository
github
Last release
2 years ago

Archived

This plugin has the potential to hang for a long time during export of large log groups. View our new plugin serverless-log-keeper which leverages DeletionPolicy: Retain to actually keep log groups after lambda deletion instead.

@xoi/serverless-log-dumpster

Serverless plugin that archives CloudWatch Log Groups before their deletion

npm Version badge Unit Tests codecov

Usage

  1. To use this serverless plugin, you must first install it as a dependency.

    Run npm install --save-dev @xoi/serverless-log-dumpster or yarn add --dev @xoi/serverless-log-dumpster

  2. Create an S3 bucket in the same region in which you are deploying your project. Note that your bucket must include a bucket policy to allow CloudWatch to write to it. You may refer to the official AWS documentation on how to do so or view the bucket policy below

    {
      "Version": "2012-10-17",
      "Statement": [
        {
          "Action": "s3:GetBucketAcl",
          "Effect": "Allow",
          "Resource": "arn:aws:s3:::<REPLACE_WITH_BUCKET_NAME>",
          "Principal": { "Service": "logs.<REPLACE_WITH_REGION>.amazonaws.com" }
        },
        {
          "Action": "s3:PutObject",
          "Effect": "Allow",
          "Resource": "arn:aws:s3:::<REPLACE_WITH_BUCKET_NAME>/[OPTIONAL_PATH_PREFIX/]*",
          "Condition": { "StringEquals": { "s3:x-amz-acl": "bucket-owner-full-control" } },
          "Principal": { "Service": "logs.<REPLACE_WITH_REGION>.amazonaws.com" }
        }
      ]
    }
  3. Add the plugin to your serverless.yml and add its configuration properties

    [...]
    region: us-east-2
    
    plugins:
      - '@xoi/serverless-log-dumpster'
    
    logDumpster:
      destinationBucketName: example-logdumpster-bucket
      # destinationPathPrefix: defaults to 'logdumpster'
    [...]
  4. Continue using Serverless as you normally would! LogDumpster will take care of creating and awaiting log group export tasks for any log groups that will be removed as a result of updating the cloudformation stack.

  5. In the event of plugin or bucket misconfiguration, LogDumpster will error out and prevent the deployment from continuing keeping your log groups intact.

Configuration

All configuration is done in serverless.yml at the root level under the logDumpster property.

Here are the possible configuration options: | Configuration Key | Required? | Default | Description | | ----------------- | --------- | ------- | ------------| | destinationBucketName | Yes | N/A | Name of the AWS S3 bucket to export logs to. Please note that logDumpster will not automatically create nor manage this bucket for you. Refer to the Usage section of the README for more information on how to configure this. | | destinationPathPrefix | No | logdumpster | S3 path prefix for export tasks. A slash is automatically appended to the end so that the exports live in s3://thebucket/prefix/log-group-name/<data>

Other useful information

LogDumpster looks at the raw CloudFormation template diff (deployed vs deploying) using @aws-cdk/cloudformation-diff for any deletion, changes or replacement to AWS::Logs::LogGroup resources, regardless of their origin. This means that it will export log groups before deletion for any log group specified within the CloudFormation template, even if it was created and deleted manually or is managed by another plugin.

Development and testing

We aim to keep the core of the plugin purely functional to allow great unit test coverage. However, you might want to test the plugin directly with Serverless. For this purpose, two "demo" projects exist: demo-projects/create-log-group/ and demo-projects/delete-log-group.

Helper npm scripts exist in package.json, namely: demo:create, demo:log, demo:delete, demo:teardown and finally, just running yarn demo will run all four sequentially.

Note that these scripts use your OS username ($(whoami)) as a suffix for the bucket. You must create the bucket yourself and configure it as specified above.

Contributing

We welcome all contributors with open arms! See CONTRIBUTING.md