I’m fairly confident, unless I missed something that server memory allocation isn’t being handled properly for our servers, that roughly contain ~100 players on average.
Our servers seem to be crashing at 8gb of memory, due to the high player counts memory easily can rack up in our complex game. Following the formula and update posted in late 2024.
We should be roughly getting 6.4GB + 100MB * ~105 peak = ~17gb of server memory. I also wonder if the CPU Allocation (Although stated for certain experiences), works.
Hey @Dev_Badge2. Could you provide the place id in question? It’d help with the investigation.
For your context, in the 6.4GB + 100MB * peak_players formula, peak_players refers to the maximum number of players this specific RCC instance has seen since its start. May it be the case that your servers go out of memory before that many players join?
Thanks for providing the necessary details. I checked the telemetry and didn’t find a lot of OOM terminations for the place id you shared. Also, I verified that the memory limit was recalculated correctly for instances of this place, keeping up with the number of players as they join in. If you still encounter problems, additional context would be helpful. For instance, if you see an instance terminated, any details that would help to identify it would be appreciated. Such as date and time of start and/or termination, known player ids who joined this server, etc.
Our most recent server crashed at approximately 1:17 PM CST 7/7/2025
The server had reached a peak player count of at least 85 players before hard-crashing
The server crashed roughly around the same range of memory as referenced in Hyperant’s bug report.
@Dev_Badge2
I pulled up logs for that RCC instance. There was a crash, though; there’s no evidence that it was due to memory consumption, as the memory limit was correctly updated. The bad news is that it’s not easy to determine why it crashed. I’ll see if I can get more clarity on that.
@Hyperant
Feel free to send me the place id in question. I pull up telemetry for your place as well. Any details (like crash date/time and player occupancy at that moment) would help the investigation.
I deactivated the server limit on this particular server just a few minutes ago and sure enough, it crashed around 24 players at just below 8100 mb server memory
It turned out that in some cases, a different tool limited memory consumption on another level, which is why I didn’t see any issues on the engine level. I deployed a configuration change that should have addressed that (only for RCC instances started after 3 pm PST today). Please let me know if you still see the problem with the new instances.
