it looks like there is a bug on deployment of serverless endpoints. testing takes too long and it fails with 2 hour timeout of testing time. the solution is very easy, just delete the test file from .runpod/tests.json file. that's it. though it says writing test is required in the setup section, it doesn't. it just skips the testing point and your deployment turn green and become completed. thats it 🙂
#testing takes too long and fails > the solution is here!
8 messages · Page 1 of 1 (latest)
Hello, I have a question. If I use rtx A5000 gpu which shows 0.00008/s, does that mean I will be charged only when I am using it??
When a worker is running it will charge that much per second (as shown in the pricing)
Ok thank you. How much time will it take for the gpu to become idle after some prompts say i am using it continuously for 20 minutes after which I stop chatting with an ai model. After this how long will it take for the server less gpu to become idle?