do we get billed partially or rounded up to the second?
If my execution time is 0.35 seconds, will I get billed 1 second for that request or partially?
3 Replies
also are we only getting billed for executionTime or executime+delayTime?
Delay time should not be billed as long as it's RunPod's delay (e.g., a job waiting in the queue) and not your code (cold start). It's good practice to put app initialization, such as loading AI models into VRAM, outside the RunPod serverless handler function. This is then marked as a delay in stats yet still billed as execution time.
What is billed is Cold start time + execution time, so when the worker is running its billed, either loading model into vram or running and yes it's rounded up per second