Every job you run spins up one or more containers: we keep track of the container usage—and not the machine capacity—to fairly establish the price of your job. This includes recording information about the vCPU and RAM utilization of the containers over their lifespan.
We use a simple metric called ByteChip to record computing usage. It's equivalent to one hour of compute on a unit container:
C1 = 1 hour computing on 1 vCPU + 1 GiB
This metric makes it easier to track compute time. For example, suppose a job using 0.45 vCPU-hours and 0.89 GiB-hours costs BYC 0.49351461. This would be the cost of running a job for half an hour on a container with 1 vCPU and 1 GiB of memory.
This is the current USD value of 1 ByteChip:
ByteChips charges accumulate in your wallet, allowing you to track total computing hours. Don't worry, dollar amounts are displayed on the billing page and invoices as well, so you can always see how much you're spending.
Additionally, unlike other services, we do not charge for idle machine capacity allocation. Each app you build comes with some guaranteed idle capacity, minimizing cold starts, so that you only pay when your job is running, and not before or after.
