Hello there!
Last week we experienced sudden connectivity issues between our client apps and our api apps. The requests kept timing out until we switched the api url to Fly internal DNS. What could have caused this?
The client apps had been running using fly default domain for months. They stopped working suddenly on September the 20th.
Thanks!
Just a wild guess: The TLS certs (verify on crt.sh) of the domain pointing to Fly instances were not renewed on time? See also: How to use fly.io to manage the multiple domains for our project - #6 by jerome
If I understand correctly, @ilrock is talking about their .fly.dev hostnames.
Can you tell us which app is showing this issue? I see a few apps on your account, knowing with one (or a hint as to which one) would help us track this down.
I might’ve found you app. Looks like there are no A or AAAA records for it. Looking at your app, it has no IP assignments.
Were those removed on the 20th?
Hey Jerome! So it happened on several apps following this naming pattern “site-prod-XXXXX” and “site-stage-XXXX”. They were trying to connect to the two API apps we have “api-prod” and “api-stage”.
Both api apps do have IP addresses assigned to them.