Production insight

The catalog tells you what your solution should do. These four areas, unlocked by uploading server logs, tell you what it actually does. No other FileMaker analysis tool reads production logs at all, so this section may take a moment to trust: everything here comes from your server's own records.

Server Logs

Observed production activity: what actually runs, how often, and what breaks. This is the ground floor the other three areas build on, and the place to confirm the ingest picked up what you expected.

Performance

Server health and where server time actually goes, built in part from FileMaker Server's TopCallStats.

Schedules

Every server schedule on a zoomable timeline: when each one ran, how long it took, what overlapped, and what's coming up on the calendar. Schedule problems are usually shape problems (two heavy schedules colliding at 2am), and shapes need a picture.

Incidents

The reconstruction view. Pick a moment ("things went sideways around 2:30pm on the 14th") and see what was happening on the server around it: which scripts, which schedules, which errors, which sessions. Built for the morning after.

Production evidence, everywhere else

Once logs are in, evidence shows up where you're already looking:

• Entity pages gain observed-activity panels: does this actually run in production?
• Unused elements gains a server-side usage view.
• The AI tools for errors, expensive calls, schedule health, incidents, and server health come alive.

Two honesty rules

1. Coverage is always stated. Every log-backed claim carries its window ("based on logs covering March 3 to June 12"). If a period isn't covered, FM Fathom says nothing rather than implying a clean bill of health.
2. Your error policy applies. Codes you've marked non-fatal in the Error Catalog are treated that way in every report here, so known noise doesn't bury real signal.

Related

• Upload server logs
• Prompt recipes: asking the AI about all of this in English