Open dmumpuu opened 9 months ago
Edit: It's even worse, the whole thing now cannot be destroyed with the Terraform CLI. That means I gotta manually go in and delete the resources. π
The examples are outdated and don't work. I tried the eks_argo TF example, and it would bork out with
β Warning: Argument is deprecated
β
β with module.metaflow-datastore.aws_s3_bucket.this,
β on .terraform/modules/metaflow-datastore/modules/datastore/s3.tf line 1, in resource "aws_s3_bucket" "this":
β 1: resource "aws_s3_bucket" "this" {
β
β Use the aws_s3_bucket_server_side_encryption_configuration resource instead
β
β (and one more similar warning elsewhere)
β΅
β·
β Error: creating Lambda Function (metaflowdb_migrateir9nhhph): operation error Lambda: CreateFunction, https response error StatusCode: 400, RequestID: XXX, InvalidParameterValueException: The runtime parameter of python3.7 is no longer supported for creating or updating AWS Lambda functions. We recommend you use the new runtime (python3.12) while creating or updating functions.
β
β with module.metaflow-metadata-service.aws_lambda_function.db_migrate_lambda,
β on .terraform/modules/metaflow-metadata-service/modules/metadata-service/lambda.tf line 115, in resource "aws_lambda_function" "db_migrate_lambda":
β 115: resource "aws_lambda_function" "db_migrate_lambda" {
Also, the EKS version in the examples is outdated and is unsupported from March 2024. I understand that these are simply some first steps, but still a bit disappointing. We built our own Metaflow deployment with AWS CDK but CDK got its own issues and AWS step functions is excruciatingly slow, so I was really hoping for some speed improvements using Argo + k8s + TF both for deployment of the infrastructure and for deployment of workflows.
Steps to reproduce:
cd terraform-aws-metaflow/examples/minimal
locals.resource_prefix = "test-metaflow"
in minimal_example.tfterraform apply
and wait until it finishesaws apigateway get-api-key --api-key <api-key> --include-value | grep value
and paste the result to themetaflow_profile.json
filemetaflow configure import metaflow_profile.json
python mftest.py run
mftest.py
The running task will never finish, created AWS Batch Job in the AWS Batch Job queue is always in status
RUNNABLE
Also tried with
outerbounds/metaflow/aws
version=0.10.1 andterraform-aws-modules/vpc/aws
version = 5.1.2Generated metaflow_profile.json
```json { "METAFLOW_BATCH_JOB_QUEUE": "arn:aws:batch: