Resource Monitor

Find out which resources cost you frame time, using FXServer’s own profiler.

How it works

This is an on-demand capture, not a live readout. Open the Monitor tab in the bottom panel and click Profiler Snapshot. Boltise tells the running server to record 300 ticks — roughly five seconds at 60fps — then saves and parses the trace. The whole round trip takes about nine seconds.

You get a table of every resource that ran during the capture:

Column	Meaning
Total ms	Script time the resource consumed across the whole capture
Avg ms	Its average per tick
Worst tick	Its single most expensive tick
Calls	How many times it ran
% of total	Its share of all script time in the capture

Above the table is a summary of the capture as a whole: total frames, average script time per tick, worst and 95th-percentile frame times, how many ticks ran heavy, and a hitch count.

Note

The profiler measures script time. It does not report per-resource CPU percentage or memory — FXServer does not expose those per resource. Process-wide status and tick budgets are on the same Monitor tab; see Server Management.

Requirements

The profiler only works against a server Boltise itself started, because it reads the trace file the server writes to disk. If you are attached to a remote server, the snapshot button reports that there is nothing to profile.

Only one capture can run at a time — two overlapping recordings interleave into a trace that describes neither.

Snapshot history

Every capture is kept in .boltise/profiles/ inside your server-data folder. Profiler History lists them and reloads any one, so you can record before and after a change and compare the two.

Finding bottlenecks

1. Start your server from Boltise and let it reach a steady state — ideally with players on, since an empty server profiles nothing interesting.
2. Open the Monitor tab and click Profiler Snapshot.
3. Sort by % of total to find the resources that dominate, then by worst tick to find the ones that stutter.

A resource with high total time is a steady cost; a resource with a low total but a huge worst tick is what players feel as a freeze.

Common culprits:

• loops without a proper Wait()
• heavy database queries running every frame
• work done on every tick that only needs to happen on an event

You can also ask the AI panel to take a snapshot and read it for you — it has the same capture available as a tool and will name the offenders directly.