Hi, nice summary you have here.
Few things i've noticed:
Max timeout is more refined than 15 min. it's up to 1h on managed cloud run (though pre-ga for now) and i think up to 24h on anthos.
On billing you say If your service handles multiple requests simultaneously, you do not pay for them separately this is only partially correct. With concurrent requests you don't pay for additional cpu/ram, but you pay for each request individually.
A note that with anthos you don't pay per request might help showcase the best pricing model: cloudrun on gke on premptible might be the sweet-spot of computing today.
Hi, nice summary you have here. Few things i've noticed:
If your service handles multiple requests simultaneously, you do not pay for them separately
this is only partially correct. With concurrent requests you don't pay for additional cpu/ram, but you pay for each request individually.