Problem

A running StackState instance can become slow, unresponsive or accumulate too much data on the disk. This can have multiple causes, here we list the tools at our disposal to gather and interpret statistics about a running StackState instance.

Cache metrics

The file var/log/transaction.metrics.log contains a detailed overview of transactions that have run and how many cache hits/misses/data volume occurred. When views are slow, this is the first log to check. When getting too many cache misses, consider enlarging the cache.

Storage statistics

StackState can become slow or retain a lot of data (which in turn makes it slow). Storage statistics help out to get a breakdown of how much data is currently in StackState, how much was in history: all broken down for the various domain objects to home in on the problem. To gather storage statistics, use following command:

echo "storageStatistics()" | stackgraph/bin/console.sh > storage.txt