This article applies to PRTG Network Monitor 14 or later

Reasons for High Memory Usage on PRTG Servers

There can be various reasons for high memory consumption on servers where PRTG probes are running. Most likely, this comes from over-utilization of PRTG, the general system setup, or interfering Windows programs. Usually, PRTG does not leak memory but in some rare cases there might be system internal issues causing high memory consumption by PRTG itself. If none of the points below apply to your setup, please contact our support via the support bundle.

This article sums up various parameters you should check if you encounter high memory values:

• Check the PRTG system requirements. Minimum requirement for RAM memory is 1024 MB for a common usage scenario of PRTG. The bigger your installation, the more RAM you should provide. Also remember that 64bit systems need more memory than 32bit systems!
• Sensor related:
  • Too many sensors for the underlying system resources: One sensor needs approximately 150 KB RAM.
  • Too many channels. The more channels a sensor has, the higher its memory usage.
  • Too many sensors with high system impact.
  • Too high values for the graph cache. For details, please see PRTG Manual: System Administration—User interface, section Graph Settings.
  • Check the page PRTG Status—System Status in your PRTG web interface for system analysis. On this page, you can find the total number of sensors and channels in your installation, sensors by their performance impact, and information about core system memory.
• Toplist related:
  • Your PRTG installation contains too many toplists.
  • Toplists with very heterogeneous traffic and long intervals consume a lot of RAM. Minimize the covered time period in the settings of a toplist.
  • Toplists with too many entries per toplist need a lot of memory. Decrease the Top Count in the settings of a toplist.
• Overall system related:
  • Too many users which are active at the same time on one PRTG server can cause many sessions and states which then again lead to increasing memory usage.
  • Probes that try to re-connect very often cause high memory usage. Ensure a stable network connection for your remote probes!
  • Check the system(s) on which a PRTG probe is running for interfering Windows programs and services, for example, via the task manager.
    • For example, a customer found out that the Microsoft Network Store Interface (NSI) service leaks memory. For details and hotfixes, see the Microsoft articles A memory leak in NSI Service occurs unexpectedly and Svchost.exe running NSI service leaks memory and non-paged pool memory leak.
  • If necessary, stop service(s) with heavy load and clear the cache on the system running your PRTG probe.
  • Always run the latest PRTG version! Outdated PRTG versions might contain unsolved issues. These can result in an unnecessary high number of CPU threads which then again round off RAM.
  • Check the memory consumption on your PRTG core server system: Choose Setup | PRTG Status from the main menu in the PRTG web interface and navigate to section Core System Memory where PRTG tracks all important memory areas. There you might find potential reasons for memory leaks.
  • Check also your Core Health and Probe Health sensors to find out which system values increase.