1 critical, 2 warnings found. Every finding below comes from what was actually in the pasted files. Treat the fixes as suggestions: they cover the common causes, and this case may need more digging.
Neoforge
What this error means, and every way to fix it →
tpa read its config before it was loaded
tpa 2.0.0 asked for a config value before the loader had read that config file, and the loader refuses rather than hand back a guess. Server configs are not loaded until a world is, so a mod that reads one from the title screen or during startup crashes like this. This is a bug in tpa, not in your setup. Update it, since fixes for this are usually quick to ship. If its config file in config/ or serverconfig/ is empty or broken, deleting it lets tpa write a fresh one. If neither helps, remove tpa and report the crash to its author with this report.
Usual suspects: tpa.
which explains: The server crashed and then hung
The hang is the aftermath; the crash above is the story.
AllTheLeaks is reporting objects the server cannot free
The AllTheLeaks mod is printing its leak report: LevelChunk (minecraft): 2184, ServerPlayer (minecraft): 6. Invalid cloned players: 0. Memory in use grew by 2991MB between its checks. Objects the game has finished with are still referenced by something, so garbage collection cannot reclaim them and the heap fills until the server stalls or dies with an out-of-memory error. The bracketed name after each leaking type is the mod holding the reference; update or remove those first, minecraft itself rarely leaks. Cloned players point at a respawn or dimension-travel bug in a mod. Until it is fixed, restart on a schedule before the heap fills, and run /atl force_refresh after players log out to see which mod keeps growing.
jump to line 101The tick loop fell 10261ms behind, 205 ticks in one go
A tick took far longer than the 50ms it is allowed, so the server ran behind real time and skipped ahead to catch up (58 times in this log). Players see rubber-banding, delayed block breaks and mobs that stutter. One burst after a restart, a world upgrade or a chunk pre-generation run is normal. A steady stream is lag with a cause: profile it with spark (/spark profiler start --timeout 120 in the console) rather than guessing. The usual culprits are view-distance and simulation-distance set too high, entity or item build-ups, chunks generating live as players explore, and garbage collection pauses when the heap is too small.
which explains: Players are timing out mid-session
Players drop because the server stalls past the keepalive window, not because of their own connection.
2197 sensitive values (addresses, passwords and the like) were stripped before this page was stored.
Showing 2000 of 3111 lines, after redaction: the start, the end, and the lines the findings point at.
Need more help or want to discuss your issue further? Join our Discord and bring the link to this page.
Join our DiscordRun a community? Add the Roastd bot and anyone can roast a crash log in your support channel.
Add the Roastd botGot a server of your own that deserves a second opinion?
Roast your own serverRoastd is a free tool built by Hostd. This diagnosis stays put. The log it was read from is deleted on a timer.