Hey folks!
We’ve made some improvements on the usage page.
Now it’s easier to see the total amount used by each type of machine (for apps v2) and allocations (for apps v1). They’re also broken down by app.
Same thing also works within an app:
We’re looking for other ways to improve the usage page - what else would you like to see on those pages?
It’s looking great although I think the usage numbers are wrong?
I have this app that I created 8 days ago and the usage report says:
1 month, 1 day, 4 hours, 5 minutes
Another app which doesn’t exist anymore and only ran for a couple of days and it says:
2 months, 1 week, 5 days, 22 hours, 36 minutes
And yet another app has been running for 8 days and it says:
5 hours, 12 minutes
I saw that you added the “(deleted X days ago)” which is great.
An improvement you could make is having the CPU usage seconds aligned to the right (instead of left) and having a visual separator for thousands. Like in accounting documents (invoices, etc).
Right now it’s difficult to parse these numbers and compare them.
Hey Pier!
Yeap-- we’re working on making those pages easier to understand. Thanks for your suggestion, I’ll take a look!
What else have you been looking for re: usage?
This is great! Will this data be accessible via api? I’d like to get usage for a machine based on a time range programmatically.
cool-- that would be nice, but we don’t have an ETA yet. Is there a specific problem you’d be able to fix with that api endpoint?
I have a hosting service, and so I’m charging per hour of uptime for each app. I’d like to query the run time for each app given a time range.
@yroc is the idea here to show it to your users for their app? How would you want to use this data if you had it?
This is for me to compute the monthly usage bill for the user.
To summarize, at the end of the month, Stripe sends a webhook saying the user’s invoice was created. I’d then query fly.io for each app’s uptime inside the invoice’s time range. I’d then update the invoice with the returned uptime. Stripe would multiply the uptime by the rate I charge per hour for the product.
Is that clear?
