I’m not sure if I’m calling the workflow incorrectly or what. I followed documentation about sending about a dozen “warm up requests” with tiny images to “warm up” the serverless API endpoint…but almost half of my requests end up being 502 which is causing a bad experience for my users.
Is this just the nature of using serverless? Am I doing something wrong?
Good morning @Austen_Payan,
My name is Ford and I am a Support Engineer at Roboflow. I’m incredibly sorry for the errors you have experienced today, and especially for the downstream impact on your customers. This is not the experience you deserve, and we’re committed to making this right.
To help me triage further, do you grant Roboflow Support Permission to access your workspace? Additionally, can you please provide the name of your erroring workflow?
Thank you!
Yes I give permission. The workflow in particular is canaveral/sahi-yolo-2, however I have experienced this on many different workflows and individual testing of models on their own outside of a workflow (all via the serverless mechanism)
Good afternoon @Austen_Payan,
Thank you for the additional context! Quick update here, unfortunately I was unable to replicate this 502 error running inference on your workflow or directly on your model.
To help me triage further, can you please provide a sample code snippet you’re using to run inference and a sample testing image that is consistently throwing this error via Google Drive link to maintain image quality?
Please feel free to direct message the code snippet or link to images to maintain privacy.
Thank you!
Hi Ford, apologies, how do I direct message you?
Good morning @Austen_Payan!
No worries at all! To send a direct message, click on my Profile Icon in this thread. That will take you to a Profile page with a “Message” button in the top right corner, which you can use to send a Direct Message.
I have reached out to your account via Direct Message, you should see a notification. Thank you!