
When should automatic Rust recovery send you an alert?
Write down the alert you’d want if the game stays down and the note you’d want if it comes back by itself. Then test both routes. A short interruption that ends with players able to join may only need a record. A world players still can’t enter needs someone to check it.
Compare a recovery notice with a failed recovery
Use your monitoring tool’s test mode or a spare server to check whether a recovered problem and an ongoing outage reach different destinations. One example is to keep ‘Evening Builds recovered’ in a log channel and send ‘Evening Builds still can’t accept joins’ to whoever agreed to cover that evening. Ask the recipient to name the first thing they would check.
LinuxGSM’s monitor is an example of a tool that can check a game and attempt a restart. Read your own tool’s event definitions. A restart attempt and a player being able to join again are different observations.
Give repeat notices an owner
Check what happens if the same spare game stays down through the next check. If the host and management tool both send a ping about the same outage, choose one alert the helper will act on and keep the other as history where the tools allow it. Confirm the alert includes the world name and enough context for the helper to try a join or check the game without opening every server.