Send useful diagnostics.
Spark profiles are best for performance triage. Latest logs, crash reports, debug logs, JSON, text files, and zipped diagnostic bundles can help too.
Use this guide to upload the right files, read the important sections first, and compare fixes without guessing your way through server performance.
Spark profiles are best for performance triage. Latest logs, crash reports, debug logs, JSON, text files, and zipped diagnostic bundles can help too.
Read the plain-English summary, grade, top systems, top mods, and likely causes before digging into raw hotspots.
Do not upload private keys, payment exports, passwords, tokens, database dumps, or unrelated personal data.
Capture one profile during real load, apply one fix, then capture again. Comparison only helps when the test is clean.
Capture while the server is doing real work.
Change one thing so the result means something.
Compare grade, MSPT, TPS, and recommendations.
Use a Spark profile captured while the issue is happening. If the problem is lag, tick time, or low TPS, this is usually the most useful starting point.
No. AI is an optional explanation layer built from TickVisor's parsed report data. The normal report still works without AI credits.
Comparison shows whether a change improved or worsened the server instead of relying on vibes, the worst metric ever invented.
Upload logs when the profile alone does not explain a crash, startup failure, warning spam, or mod-specific error.