Mm weird, because my caddy proxy just randomly dies suspiciously close to when the error and warn logs start appearing. It might be something else, but it only started happening after I pulled the new 2.8.x versions.
Do you have any details? What are the logs before it happens? What do you mean it “randomly dies”?
Sorry, my bad, I made that comment in a bit of a hurry.
Since the update, I started noticing that every once in a while (multiple times a day), my websites stopped being reachable. When looking at the caddy logs, there was nothing other than ARI errors and warns, like these:
WARN tls.cache.maintenance ARI window or selected renewal time changed {"identifiers": ["my.domain.tld"], "cert_hash": "...", "ari_unique_id": "...", "cert_expiry": "...", "prev_start": "...", "next_start": "0001/01/01 00:00:00.000", "prev_end": "...", "next_end": "0001/01/01 00:00:00.000", "prev_selected_time": "...", "next_selected_time": "0001/01/01 00:00:00.000", "explanation_url": ""}
ERROR tls.cache.maintenance updating ARI {"error": "could not fully update ACME renewal info: either no ACME issuer configured for certificate, or all failed (make sure the ACME CA that issued the certificate is configured)"}
A simple restart solved the issue, so I ended up writing a script that curls an endpoint and if it fails, the container is restarted.
I can get more details if needed.
Can you pull a profile when it’s not responding?
A goroutine profile would be most helpful.
Will try! It may be a while before I get one, because it’s very unpredictable as far as I can see.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.