Direct answer
Lag after a Minecraft or plugin update may come from one-time conversion work, a changed plugin, new world generation, different defaults or a genuine regression. Separate transient startup costs from repeatable tick problems.
How to check it
- Record exact before/after game, server, Java and plugin versions.
- Watch startup logs for migration warnings and distinguish first launch from repeated restarts.
- Capture a spark profile during the same workload that was previously healthy.
- Compare player count, view/simulation distance, mods and active worlds so the test is fair.
What to change
- Test new chunk generation separately from already-generated regions; a world update can create heavy terrain-generation work.
- If one plugin appears in the hottest stacks, check its release notes and test a compatible build in staging.
- If the update is incompatible, restore a complete pre-update world/plugin backup; do not downgrade a converted world in place.
Software and version notes
Some software changes config defaults or deprecates keys. Blindly copying old tuning values forward can conceal new behavior or cause an unintended workload increase.
Practical example: Example: lag visible only during the first post-upgrade boot may be registry conversion or cache rebuilding. If peak-hour MSPT stays high for days, collect a profile with actual players and investigate persistent workload changes. A version upgrade and a plugin upgrade done together must be separated on staging to identify the culprit.
Verify the result
- Repeat the profile once initial migration finishes.
- Validate affected player features and world data in addition to MSPT.