Downloading from a private S3 bucket

General
1

Hello! I have two questions about using a private S3 bucket for artifacts as described in these docs.

First: the docs seem to describe two ways of specifying a private S3 bucket during an upload:

# Supply it as a target destination
buildkite-agent artifact upload "file.log" s3://name-of-your-s3-bucket/

# Or define a destination env
export BUILDKITE_ARTIFACT_UPLOAD_DESTINATION=s3://name-of-your-s3-bucket/
buildkite-agent artifact upload "file.log"

Is there any difference between those two approaches?

Second: if an artifact has been uploaded to a private S3 bucket using one of the strategies above, will it work out of the box to do buildkite-agent artifact download “file.log” in a later job? In other words, will buildkite-agent remember the file was uploaded to our private S3 bucket and find it, or does it need some special handling?

Thanks for any help! I’m working in an environment where I don’t have permissions to test this directly, so I’m just trying to be as sure as I can about the behavior.

2

Hey @bdhamilton! I can help with this. First, there is no difference between those two methods of uploading artifacts to S3. I’d say that using the s3:// path in your buildkite-agent artifact upload command is good for one-off uploads to S3, whereas setting the BUILDKITE_ARTIFACT_UPLOAD_DESTINATION env var is better when you want all artifacts to go to S3 - you can just set that globally on all your agents.

When downloading, you don’t have to specify a source path for the artifact - Buildkite will point the client to the right place based on the build number and artifact you are asking to download.

Please let us know if you have any other questions!

3

Thank you so much, Owen! I appreciate it.