This article applies to PRTG Network Monitor 19 or later
Error because of Dropped Flows (PE111)
Reasons for PE111
Error PE111 happens if PRTG is not able to process the incoming flows fast enough. To process a flow, PRTG first has to check the include and exclude filter and then all the rules used for the channels.
Depending on the number of flows per second, the complexity of the rules and the processing power of the probe system, the error might then be triggered sooner or later.
Suggestions for Troubleshooting
Try the following to resolve the issue:
- Make sure that no other process is lowering system performance.
Such processes can be scheduled tasks (like backups or virus scans) overloading the probe at specific times, or events like a search engine scan on your web server.
The System Health sensor can help here. Create a Windows Process sensor if you suspect a specific process.
- Check the complexity and number of rules that you use.
If you use very long filter expressions, try to optimize them for faster evaluation. Put the most likely matches first, use brackets to structure the rule, and use IP ranges and masks instead of separate IPs.
- Check your toplists.
Toplists need a lot of processing power, too, and can block the core-probe communication. Delete all toplists that you do not necessarily need. Toplists with fewer fields are preferable. Disable "live" toplists where they are not required.
- Use a remote probe on a separate machine.
Best performance can be achieved with a physical (non-virtual) system with 4 (or more) cores. Each sensor has its own flow processor using a thread, so more cores help if you have a lot of Flow sensors.
- Use multiple remote probes.
Distribute the load among multiple machines. If you have multiple routers, send the flow streams to different probes. If you have one router, you can send the same stream to two probes with only half the number of sensors.
Add comment