I have a bucket whose objects have commitId as their name. I want to pass these commitID to my codepipeline and utilise those id for slack messages.

I am trying to trigger Codepipeline when a zip file is uploaded to s3, however as I can see in the documents, it can only trigger with a static bucket key. I want to trigger with any file name

https://docs.aws.amazon.com/codepipeline/latest/userguide/action-reference-S3.html I am dealing with a use case where the uploaded object in s3 will have dynamic object keys.

How to deal with this situation?

I have read this question so I know using s3 with lambda and then trigger pipeline with lambda but this will still not work because I need to pass zip file to codebuild

CodePudding user response：

TL;DR Have the Lambda record the ID in commit_id.txt and add it to the bundle.

I understand you want to execute a pipeline when an arbitrary object, say a5bf8c1.zip is added to a S3 path, say MyPipelineBucket/commits/. The pipeline has a S3 source, say MyPipelineBucket/source.zip. Your pipeline executions also require the file name value (a5bf8c1).

1. Set up S3 Event Notifications on the bucket. Apply object key name filtering on the MyPipelineBucket/commits/ prefix.
2. Set a Lambda Function as the destination
3. The Lambda receives the Commit ID in the event notification payload as the triggering file name. Write it to commit_id.txt file. Using the SDK, get the MyPipelineBucket/commits/a5bf8c1.zip bundle from S3. Add commit_id.txt to the bundle. Put the new bundle to MyPipelineBucket/source.zip. This will trigger an execution.
4. In your pipeline, your CodeBuild commands now have access to the Commit ID. For instance, you can set the Commit ID as an environment variable:

COMMIT_ID=$(cat commit_id.txt)
echo COMMIT_ID # -> a5bf8c1