Find what’s slowing FileMaker down with Genie and server logs

1. Download the logs for the affected period

In FileMaker Server Admin Console, open Logs → Log Viewer. In the version shown here, Download All Logs retrieves the logs in one action. Keep the archived files that cover the period when the slowdowns occurred, too.

This investigation includes TopCallStats.log, Stats.log, ClientStats.log, Event.log, script logs, and Data API logs. Start with the files available and specify when the problem happens. If TopCallStats is missing, its collection needs to be enabled for a new measurement period.

Download a copy of the logs covering the period you want the agent to investigate.

2. Add the logs and connect Genie through MCP

Add the downloaded files to your agent’s conversation or working folder. Connect Genie through MCP and select the FileMaker file that corresponds to the logs in Genie.

That connection lets the agent move from a technical identifier in a log to an actual field, its calculation, or a script to inspect. In this example, the agent asks for the active file to be switched before verifying the matches.

The log files and FileMaker context are available to the same investigation.

3. Ask for a report that helps you decide what to fix

Give the analysis a purpose: understand the slowdowns and identify the next checks. Ask the agent to use Genie to match observed operations to FileMaker objects, then present the findings in a readable report.

Analyze these FileMaker Server logs. Use Genie through MCP to match slow calls to tables, fields, and scripts in the active file. Build an HTML report showing the highest-cost operations, worst individual calls, affected times, and next checks. Separate confirmed findings from possible causes. Include CSV rankings for fields and scripts where the logs allow them to be identified.

The agent combines log analysis with metadata available through Genie.

4. Focus on the slowdowns your team experiences

The first results may highlight an expensive overnight job. If your problem affects daytime users, say so. You can exclude a known job, a time window, or a test environment and have the agent recalculate the rankings.

Here, night scripts are excluded and the analysis focuses on weekdays from 8 a.m. to 6 p.m., Toronto time. That window is a choice for this investigation; use the hours when your team experiences the problem.

Exclude night scripts from this analysis. Focus the report on weekdays from 8 a.m. to 6 p.m., Toronto time, and state those exclusions in the report.

A clarification in the conversation focuses the report on the problem you need to solve.

5. Turn the rankings into specific leads

The report at the top of this article ranks operations and fields by observed time in the sampled calls. In this example, finds account for about 62.8% of that time. Two fields at the top of the ranking are already indexed according to the context returned by Genie, directing the next investigation toward the find criteria and the workflow using them.

The report also identifies a 320.6-second find associated with an unstored calculation. These are specific starting points for investigation. The screenshots show findings and recommendations; they do not yet demonstrate a performance improvement after a fix.

6. Ask where the expensive work starts

The report gives you the next question. Once the costly fields are identified, ask the agent to find where they are used. Genie can supply the context to inspect the relevant scripts, calculations, and layouts.

If the logs cannot establish the exact path of a call, ask what additional logging would help reproduce it. You now have targeted checks and a basis for comparing measurements after a change.

Where are these two fields used? Use Genie to find references in scripts, calculations, and layouts. Identify the possible paths that trigger these finds and the logging needed to confirm which one is responsible.

Explore a finding further using the report and the context of your own file.