Server statistics log

When you enable server statistical logging on the Logs > Log Settings tab, the Stats.log file is created in the FileMaker Server/Logs/ (FileMaker Server only) or Server/Logs/ (Claris Server only) folder. View these server statistics in Admin Console on the Logs > Log Viewer tab or by opening the downloaded Stats.log file in any application that can open tab-delimited text files. See Using log settings and Viewing and downloading log files.

You can also enable server statistical logging by using the CLI command fmsadmin enable serverstats. See Using the command line interface.

The attributes collected for each statistic include the current, average, low, and peak value.

Statistic Type

Description

Cache Hit %

Percentage of times FileMaker Server retrieved data from the cache (RAM) rather than the hard disk. Reading data from RAM is much more efficient than rereading it from the hard disk, so this value should be a high number, like 90 or 95. If it isn't, you can allocate more memory to FileMaker Server database cache.

Cache Unsaved %

Percentage of cache that is currently unsaved. This should be a low number, such as 0 or 5, so you won't lose data if you experience a system crash.

Disk KB/sec Read

Amount of data being read from disk.

Disk KB/sec Written

Amount of data being written to disk.

Network KB/sec In

Amount of data being transferred across the network into FileMaker Server.

Network KB/sec Out

Amount of data being transferred across the network out of FileMaker Server.

Remote Calls/sec

Number of separate remote calls received from all clients, divided by the statistics sampling interval. Note that there can be multiple remote calls to perform a single operation from the client's perspective.

Remote Calls In Progress

Number of separate remote calls received from all clients that were pending completion at the moment in time when statistics were last sampled.

Elapsed Time (µs)/call

Time to process a remote call. For example, a value of 1 million means an average of 1 second per call. If the average Elapsed Time/Call is high, but the average Wait Time/Call is low, ask clients whether the operations they are performing on the server have changed. Adding indexes for more fields or other solution changes may be required.

Wait Time (µs)/call

Time a remote call waits for other processes. For example, a value of 1 million means an average of 1 second per call.

I/O Time (µs)/call

Time a remote call waits for disk reads or writes. For example, a value of 1 million means an average of 1 second per call. If the average I/O Time/Call is high and the average Cache Hit % is low, you can increase the Database Cache Size. If the average I/O Time/Call is high and the average Cache Hit % is high, you can use additional hard drives or a RAID system on your computer.

When the Stats.log file reaches the log file size limit, it is renamed stats-old.log and a new Stats.log file is created.