502 Error - Model inference not working (my-first-project-3-jlvee/15)

Search before asking

• I have searched the Inference issues and found no similar bug report.

Bug

Hi Roboflow Support,

I'm experiencing persistent 502 errors when trying to run inference on my trained model.

Details:

Workspace: elishas
Model: my-first-project-3-jlvee/15 (also tried version 16)
API Key: TfznwoJ15qc17JhhRhIy
Error: "Internal error. Request to Roboflow API failed." (HTTP 502)
What I've tried:

Direct REST API call to https://detect.roboflow.com/my-first-project-3-jlvee/15
Serverless endpoint: https://serverless.roboflow.com/my-first-project-3-jlvee/15
Python inference_sdk with client.infer() - same 502 error
Created a new Workflow ("detect-count-and-visualize") - got 500 error with "Content-Length header is not numeric"
Notes:

My API key works fine (verified via https://api.roboflow.com)
I have available credits (0.53 / 30 used)
The model shows as trained with good metrics (mAP 96.9%)
This was working before, stopped suddenly
Example failing request:

curl "https://detect.roboflow.com/my-first-project-3-jlvee/15?api_key=TfznwoJ15qc17JhhRhIy"
-X POST -H "Content-Type: application/x-www-form-urlencoded"
-d "<base64_image>"
Could you please check if there's an issue with my model deployment or the inference servers?

Thank you,
Elisha

Environment

No response

Minimal Reproducible Example

No response

Additional

No response

Are you willing to submit a PR?

• Yes I'd like to help by submitting a PR!

Start by reproducing the failing requests against the detect.roboflow.com and serverless.roboflow.com endpoints shown in the issue, including the Python inference_sdk call. Inspect the resulting server-side or deployment errors; done means identifying the cause of the 502/500 responses and restoring successful inference for the reported model.