With another app where I use this library besides Guppe, https://github.com/immers-space/immers, we recently found all of the different deploys in a project unable to renew.
Similar to #6 (comment), the logs would show
| 🔒 ❨auto-encrypt❩ HTTP server is now only responding to Let’s Encrypt challenges.
| ⚠ ❨auto-encrypt❩ Received non-ACME HTTP request for /, not responding.
and then it would stay in ACME mode forever, occasionally logging out more of those "non-ACME HTTP request" warnings as stray requests came in.
One thing both projects have in common is a recent change to be moved behind a non-tls-terminating load balancer, but one was docker swarm while the other is Amazon NLB.
Interestingly, for this case we were able to resume service for now by deleting the stored certs and restarting the server, triggering a fresh cert request which completed without issue.
We're also a little behind on version, using auto-encrypt 3.1.0, because we're still on node 16.
With another app where I use this library besides Guppe, https://github.com/immers-space/immers, we recently found all of the different deploys in a project unable to renew.
Similar to https://codeberg.org/small-tech/auto-encrypt/issues/6#issuecomment-771989, the logs would show
```
| 🔒 ❨auto-encrypt❩ HTTP server is now only responding to Let’s Encrypt challenges.
| ⚠ ❨auto-encrypt❩ Received non-ACME HTTP request for /, not responding.
```
and then it would stay in ACME mode forever, occasionally logging out more of those "non-ACME HTTP request" warnings as stray requests came in.
One thing both projects have in common is a recent change to be moved behind a non-tls-terminating load balancer, but one was docker swarm while the other is Amazon NLB.
Interestingly, for this case we were able to resume service for now by deleting the stored certs and restarting the server, triggering a fresh cert request which completed without issue.
We're also a little behind on version, using auto-encrypt 3.1.0, because we're still on node 16.