Network Management

Troubleshooting polling delays in overloaded systems

Starting with SolarWinds Platform 2025.4, new notifications have been introduced to alert users when certain Job Engine (polling) counters exceed defined thresholds. These alerts indicate that polling may be falling behind due to an overloaded system.

First published date

10/21/2025 12:00 PM

Last published date

10/21/2025 12:00 PM

Overview

These notifications typically occur only in extreme cases, such as when:
  • A single poller is monitoring more than 108,000 elements.
  • The poller server is a high-performance machine with more than 16 CPU cores.
  • The system dynamically increases the supported element count per polling engine beyond the usual 108,000 limit.
Although testing shows that such loads can be manageable on powerful hardware, real-world factors—such as the complexity of polled devices—may cause polling delays.
 
Polling delays are detected by monitoring the following Job Engine counters:
  • Jobs Queue: The number of polling jobs waiting to be processed.
  • Average Execution Delay: The average delay in executing polling jobs.
If these counters exceed predefined thresholds, it indicates that the polling engine is struggling to keep up with the workload.

Product section

Orion Platform

Cause

A single engine is polling more than 108,000 elements, leading to polling delays.

Resolution

If polling delays are detected, consider the following steps to resolve the issue:
  • Reduce the load on the polling engine
    Distribute polled elements across multiple polling engines to balance the workload.
  • Increase polling engine resources
    Add CPU cores and RAM to the polling server if possible to improve performance.

If adjusting the load and resources does not resolve the polling delays, please open a support ticket for further investigation.