new pricing changes
(not relevant to me, but want to clarify)
With new pricing changes, each service gets 500 hours.
But this is easily circumvented by making another project. Is this allowed? The duplication of projects to overcome the execution limit
21 Replies
The hour limit is still at the account level not the project level
Lemme reread changelog
The difference is now we just ask a simple question “Was a workload running?” If yes, time gets consumed. If no, time doesn’t get consumed.
Wait
Ok so if I have a project with. Running deploy for 100 hours. I then delete the deploy, wait 100 more hours. Is this the difference?
But you can run as many services, and extra time won’t be consumed
If you have one deploy, it uses 100 hours at the normal rate. If you have 2 deploys, it uses 100 hours at the normal rate.
If you have anything running, it counts against your hours
But we don't "double" count it if you have 2 services running
Either you're using the platform or you're not right
Another scenario:
2 projects (A, B) running for 100 hrs, stop B and keep A running, wait 100 A, keep B running, and start tee other. How many hours do I have left? 200?
Instead of the previous 100?
your question isn’t very clear, what’s changed is that if you have anything running at all, you’ll be deducted 1 hour per hour
whether that be 1 or 20 projects
if I’m understanding correctly
Yep
does this mean the 1 vcpu and 4 gigs of ram are shared as well?
this has a lot of potential to be abused tbh
Unknown User•3y ago
Message Not Public
Sign In & Join Server To View
Yep
You're either using time or you're not
Unknown User•3y ago
Message Not Public
Sign In & Join Server To View
Next week
Unknown User•3y ago
Message Not Public
Sign In & Join Server To View
That wouldn't work for workers, or databases or anything else
So no sadly there is not
Unknown User•3y ago
Message Not Public
Sign In & Join Server To View
Yep, so if you don’t want any possible surprise charges. You can load credits.
Or you can manually crash your program and have another program monitor and send launch requests somehow?
Just an idea, I've been thinking about this problem
Well, crashed deploys count still as it still keeps a slot on the infra.
Right, but as it's crashed it'd take much less cost than a running program, right?
some overhead for the second process as well but this would be better than running a larger program