Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

Screenshot 2024-11-04 at 12.05.26 PM.png

Hi Prefect Community! A while ago we set up a flow successfully using the AWS ECS Docker setup, also using the AWS Credential prefect block. Last time we ran it, which was a few months ago, everything worked fine, but now it throws an error about the ECS config?? (Pls see screenshot or copy-paste below)  I'd be thankful for any ideas or pointers in the right direction. Many thanks :pray:
```Failed due to a(n) validation error in generating job configuration for PushPoolECSJobConfiguration. The error indicates that the input should be a valid dictionary, but it received a string instead. Check the AWS credentials configuration and ensure it is in the correct format.```

Additional info: we use a prefect `Hybrid - AWS ECS` work pool to deploy the flow.

And the log says the following:
```Failed to generate job configuration: 1 validation error for PushPoolECSJobConfiguration
aws_credentials.aws_client_parameters.config
  Input should be a valid dictionary [type=dict_type, input_value='', input_type=str]
    For further information visit <https://errors.pydantic.dev/2.9/v/dict_type>```

hi <@U01F2A93VB7>

hmm it looks like you are not using a hybrid work pool, but instead a push work pool? looking at PushPoolECSJobConfiguration

Screenshot 2024-11-04 at 12.46.37 PM.png

Screenshot 2024-11-04 at 12.44.29 PM.png

Screenshot 2024-11-04 at 12.50.35 PM.png

Hi <@U02GMEZU18B>! Thank you for your quick response!! :pray: Sorry about that, you're right, it's a `push` work pool and not a hybrid one. *I found the fix, and it seems to be related to how an empty `Botocore Config (Optional)` section in the AWS credentials block was interpreted*: I noticed a few other deployments which run on different work pools but also have an ECS setup were still functional so I went in and tried to find the relevant differences in the config. The credential blocks connected to functional ECS deployments showed `"config": null` in the `AWS Client Parameters` section, whereas the faulty credentials block would show `"config": ""`. I went into the faulty block and simply entered `null` , saved that and that fixed the issue. (Explaining this in case anybody else comes across this issue!)

thank you so much for explaining the fix! glad you got things working