Palworld Server Lag and Crashes: How to Fix Them
Most Palworld server lag and crashes come down to three things: not enough memory, memory that keeps growing until the process dies, and load from too many bases, Pals and workers. Check RAM first, watch whether it climbs over time, then reduce load with the settings Pocketpair documents. A scheduled restart is the standard practical safeguard.
Facts below come from the official Palworld server documentation (version 1.0.4 when checked in November 2026). Where something is community knowledge rather than official guidance, this guide says so.
- Pocketpair lists 16 GB RAM as the requirement and says 8 GB can boot the server but raises the chance of out-of-memory crashes.
- Admins widely report that memory climbs the longer the server runs. There is no official restart interval, so use your memory graph to pick one.
- Save, warn players, then restart on a schedule with
/Saveand/Shutdown. - Load settings to lower: bases per guild, workers per base, Pal spawn rate and the Pal sync distance.
- Keep client and server on the same game version, and store the server on an SSD.
On this page
Step 1: check you have enough hardware
The official requirements page recommends 4 or more CPU cores and lists 16 GB of RAM, noting that 8 GB can start the server but raises the possibility of out-of-memory crashes. It recommends a fast SSD and warns that slow storage may corrupt save data. The server listens on UDP port 8211 by default.
- If the process disappears with no clear error as memory approaches the limit, that pattern points to an out-of-memory kill.
- If the server is smooth empty but laggy with players, suspect CPU load from bases, Pals and workers.
Hosting with ChicagoServers? The Palworld server hosting page lists what is included, such as free server backups and SFTP file access.
Step 2: memory growth and scheduled restarts
Many admins report that Palworld server memory usage rises steadily over hours of uptime rather than levelling off. Pocketpair's documentation does not publish a restart interval or describe the cause, so treat the following as community practice, not an official rule.
- Watch the memory graph across a busy session. A steady climb toward the ceiling means you need regular restarts.
- Choose a restart schedule that keeps the server well below the ceiling, with a margin, at off-peak hours. Intervals admins report range from a few hours to daily, depending on RAM and the size of the world.
- Always save and warn players first.
/Broadcast Restarting in 5 minutes
/Save
/Shutdown 300 Server restarting, back in a minuteThe REST API, when enabled with RESTAPIEnabled=True, also has save and shutdown endpoints and uses HTTP basic authentication. The docs say it is not designed to be exposed to the internet, so use it from the same machine or a trusted network only.
Step 3: reduce the load the world creates
The official configuration reference flags these settings as affecting performance. Lower them one at a time and observe.
| Setting | What the docs say |
|---|---|
BaseCampMaxNumInGuild | Bases per guild; default 4, max 10. Higher values add processing load. |
BaseCampWorkerMaxNum | Maximum Pals per base, up to 50. Higher values add load. |
BaseCampMaxNum | Total bases allowed on the server. |
PalSpawnNumRate | Pal spawn rate. The docs note it affects performance. |
ServerReplicatePawnCullDistance | Distance in cm from players at which Pals sync; range 5000 to 15000. |
MaxBuildingLimitNum | Per-player building cap. 0 means unlimited. |
PhysicsActiveDropItemMaxNum | Maximum dropped items that use physics. |
bIsUseBackupSaveData | World backups; the docs note it increases disk load. |
Some admins also report that disabling raids with bEnableInvaderEnemy=False reduces load. That is a community observation, not documented guidance, so test it on your own server.
Step 4: startup arguments
The official arguments page lists -useperfthreads -NoAsyncLoadingThread -UseMultithreadForDS as performance-related flags and states that in v1.0 and later, leaving this parameter unset may improve performance. Many older guides tell you to add them, so removing them is worth testing if you copied an old startup line. -NumberOfWorkerThreadsServer=X sets the worker thread count and works with those flags.
Step 5: keep versions in step
Client and server need to run the same game version. After a Palworld patch, update the server and restart it; a server left on the old build can refuse updated clients. If crashes began right after you added mods or changed versions, remove the mods first and re-test.
Step 6: reading what the server tells you
Palworld's dedicated server writes events to its console output, and -logformat can switch the format to text or JSON. Whether log files appear on disk depends on how the server is run, and hosting documentation disagrees on file locations, so rely on the console history and whatever log view your host provides rather than a particular path.
- Note the last lines before a crash and compare the time with the memory graph.
- Crashes at the same moment every time, such as the same login or area, point to a bad save or a specific base.
- Crashes only after hours of uptime with climbing memory point to memory pressure.
- Crashes right after a patch point to a version mismatch or mods.
For ongoing monitoring, the REST API's /metrics endpoint returns server FPS, frame time, uptime, current and maximum players, and base camp count.
If the save is damaged
With bIsUseBackupSaveData on, the server keeps rolling backups in a backup folder inside the save data directory: 5 saves every 30 seconds, 6 every 10 minutes, 12 every hour and 7 every day. Stop the server before restoring so it doesn't write over the files you just restored. Keep a copy of important saves somewhere off the server too.
Frequently asked questions
How much RAM does a Palworld dedicated server need?
The official requirements list 16 GB. Palworld can boot on 8 GB, but the docs warn this increases the possibility of out-of-memory crashes.
Why does my Palworld server get laggy over time?
Admins widely report memory climbing as uptime increases. Watch your memory graph and restart on a schedule before it reaches the limit; there is no official interval.
Which settings reduce Palworld server lag?
Lowering BaseCampMaxNumInGuild, BaseCampWorkerMaxNum, PalSpawnNumRate and ServerReplicatePawnCullDistance, which the official docs flag as performance-related.
Should I use -useperfthreads on a Palworld server?
The official docs say that in v1.0 and later, leaving -useperfthreads, -NoAsyncLoadingThread and -UseMultithreadForDS unset may improve performance. Test with and without.