Sagemaker Endpoint Resource #1034
Conversation
Signed-off-by: Blake Romano <[email protected]>
|
|
There was a problem hiding this comment.
Thanks for the contribution! It sounds like you need a bit of help getting your local dev environment set up properly. The command to use to run codegen is simply make generate. You should also test your example using make e2e UPTEST_EXAMPLE_LIST=config/examples/sagemaker/endpoint.yaml. You'll need to install kind, kubectl, and a few other tools, and configure aws credentials in an env var as explained in a comment in the Makefile. Yes, you can make an env var with newlines in it.
Signed-off-by: Blake Romano <[email protected]>
blakeromano
requested review from
ulucinar,
sergenyalcin and
turkenf
as code owners
| matchLabels: | ||
| testing.upbound.io/example-name: example | ||
| primaryContainer: | ||
| - image: ${data.aws_account_id}.dkr.ecr.us-east-1.amazonaws.com/sagemaker-scikit-learn:latest |
There was a problem hiding this comment.
Is this the only reason for the manual intervention annotation? I'm not very familiar with Sagemaker, but couldn't you just define the endpoint and EndpointConfiguration in here, and not have a model?
There was a problem hiding this comment.
You need to have a model for an EndpointConfiguration and need an EndpointConfiguration for an Endpoint hence the need for the label
There was a problem hiding this comment.
Thank you for your efforts in this PR @blakeromano, I left a few comments for you to consider.
Please fill in how you tested the resource after all corrections as in this PR.
Signed-off-by: Blake Romano <[email protected]>
|
I believe I have fixed all the PR comments and have updated my comment with a way I was able to validate endpoint creation being successful with Crossplane locally. @turkenf |
There was a problem hiding this comment.
Many thanks for your contribution @blakeromano, I just left a small comment that you should do before merging. Since these examples will appear on the marketplace, it would be better to remove your account ID as I mentioned in the comment below.
Co-authored-by: Fatih Türken <[email protected]>
Description of your changes
Adds Sagemaker Endpoint resource.
Fixes #1029
I have:
make reviewable testto ensure this PR is ready for review.How has this code been tested
This requires an image that can be created/deleted only from the AWS CLI. In this case I use a public image from ECR of
683313688378.dkr.ecr.us-east-1.amazonaws.com/sagemaker-scikit-learn:0.23-1-cpu-py3found here: https://docs.aws.amazon.com/sagemaker/latest/dg-ecr-paths/ecr-us-east-2.html#sklearn-us-east-2.titleI update the Model resource with the full image reference for the key:
spec.forProvider.primaryContainer[0]I see the following output once I apply that and let resources create: