Memory leaks and the platform that slows down over days

[Fairman]

Memory leaks and the platform that slows down over days

Some trading terminals run fine after a restart and gradually degrade. Memory usage climbs, the interface becomes sluggish, and eventually the platform freezes or crashes. Long-running EAs, custom indicators, and accumulated chart objects are common causes.

The effect on scalping is subtle at first. Order response slows by tens of milliseconds, then hundreds. Because the degradation is gradual, it is easy to miss until a failure occurs at a critical moment.

Monitor memory and CPU usage over several days. A steadily rising trend suggests a leak. Restarting the terminal on a schedule, such as every day before the main session, is a simple mitigation.

For EAs and indicators written in-house, check for objects created but never released, arrays that grow without bound, and log files that accumulate. Third-party tools may need updates or replacement.

Automatic restart scripts and watchdogs add resilience, but they should handle open positions properly on restart.

How often is the terminal restarted, and does performance improve afterward?