I contract for teams in three time zones, so I am awake at stupid hours. If your deploy is broken at 3am your time, post it here with the error message and I will look.
Rules: the actual error, not "it does not work". What you changed last. No screenshots of text.
Taking you up on it. Container restarts every 40 seconds, exit code 137, nothing in the app log.
137 is the kernel killing it for memory. Your app log is empty because the process never gets to write. Check the memory limit on the container and what the process uses at start. I would guess a limit of 256 MB and a runtime that wants 400.
256 exactly. Raised to 512 and it has been up for ten minutes. I spent four hours on this.
Seconding the "what did you change last" rule. It is the answer nine times in ten and people leave it out ten times in ten.
Certificate renewal failed silently on a box I inherited. Renewal runs from a cron belonging to a user that was deleted.
Classic. Move it to a system timer owned by root, and add a check that alerts you when the certificate has fewer than 14 days left. Do not trust the renewal, trust the expiry date.
A cron job owned by nobody, failing quietly. I have chills.
Front-end person, scared of servers. Is it embarrassing to ask what a reverse proxy is?
Not at all. It is the receptionist. Requests arrive at one front door, and it sends each to the right room. Everything else is detail.
Ravi answered two of my customers' questions in this thread before my own support did. I should be paying you.