TPS indicates the number of ticks processed by the server per second. MSPT measures how long a tick lasts. A target of 20 TPS has a budget of approximately 50 ms per tick.
Step by step
-
Read TPS and MSPT together from the measurement tool provided by spark or server software. Take measurements when players report crashes.
-
Even when TPS is close to 20, short-term high MSPT jumps may occur. Don't just look at the long-term average.
-
If there are long ticks, check which job is currently running in the profile. If MSPT is healthy but only some players are affected, test the connection path as well.
-
After changing a setting, measure again with the same world and similar number of players. Compare the same time period with the previous report.
Keep in mind
Ping is a measure of network latency. It is different from the server's tick budget.