Sometimes the probe in my PRTG installation restarts and I do not know why. What reasons are there for these occasional restarts of the probe Windows service and/or of the entire machine running the PRTG probe?
This article applies to PRTG Network Monitor 13.2 or later
Reasons for Probe Restarts
There can be several reasons for probe restarts. Restarts can be caused by errors or they can be planned. This article provides an overview about reasons for probe restarts and gives support for unexpected restarts.
Planned Probe Restarts
There are several possibilities how probe restarts can be triggered intentionally.
- Restart command from the core: In PRTG, you can manually restart a probe. Choose Setup | System Administration | Administrative Tools from the main menu. In the Probe Administrative Tools section, you can either restart all probes (click Go! in section Restart All Probes), or restart a specific probe (click Restart Probe in the corresponding Probe [number] [name] section).
- Scheduled restarts: You can set scheduled restarts of PRTG services in the Administrative Probe Settings on the Settings tab of a probe. This causes either the probe Windows service or the entire machine running the PRTG probe to restart on the specified times. For more information, please see PRTG Manual: Probe Settings
- The probe Windows service will also restart after updates of the PRTG core server. If you have remote probes connected to your core server, they will be updated automatically after each core server update, and the probe Windows services of the remote probes will be restarted as well.
Errors Triggering a Probe Restart
PRTG is designed to perform automatic restarts of the probe Windows service if there are errors which can automatically be resolved by the service restart. If you encounter frequent unplanned probe restarts, please review PRTG’s log: From the main menu, choose Logs | System Events | Probe Related to display probe connection related log entries. Consider the Message column. The error messages are explained below:
When the connection between core and probe hangs, the probe will restart. If you encounter this issue frequently, please contact our support and send PRTG’s log files (see section below).
More Than X Unresponsive Scan Threads
When a certain number of hanging monitoring threads is reached, the probe will restart itself. You can see in the probe status files which sensors are affected. Follow the steps below:
- Choose Setup | System Administration | Administrative Tools from the PRTG main menu to create probe status files.
- In the Probe Administrative Tools section, click Go! besides Write Probe Status Files.
- Repeat writing probe status files several times with an interval of about 30 seconds.
- Find log files and open \Logs (System)\Probe State (Scheduler Debug Data) [number]. Depending on how many times you created probe status files, there is a corresponding number of these debug files.
- Open all recently created status files to compare them.
- Scroll down to the Threads section in the debug files to identify affected sensors.
- If there is a thread entry for the same sensor (see ID) in several debug files, this indicates a hanging monitoring thread. You can also identify the affected device in the corresponding thread line.
This issue especially concerns WMI and SSH sensors. Investigate which target system is concerned (see Scheduler Debug Data above). Back in the PRTG web interface or the Enterprise Console, please pause monitoring of the concerned system(s).
Perhaps you can create other sensor types for your needs and query your data using another method. For example, you can usually monitor CPU load via SNMP if you did this via WMI or SSH before. Please contact our support if you need assistance (see section below).
The probe restarts when the watchdog recognizes sensors not being processed for some time. A false reason for this error can be that the system time has changed, for example, one hour into future. In this case, the watchdog assumes that the sensors were not processed for a while and the probe restarts. This is often an issue on virtual systems. Check in the probe log file if there was a time change.
Find the log files and open \Logs (System)\PRTG Probe Log [maximum number].log to check for unusual time changes. If you experience this issue frequently, please contact our support and send PRTG’s log files (see section below).
Errors When Using Remote Probes
When using remote probes, the following errors indicate problems with your internet or network connection. Please check its stability.
- There are more than 3 access violations in connectionbase.execute
- There are more than 3 access violations in connectionbase.check
- There are more than 10 socket errors (#10093)
Log Files and Contacting Support
However, when you experience these erroneous restarts frequently, please contact our support team via the Contact Support form in the PRTG web interface and send your log files using the Support Bundle option.
Is there a way to set PRTG not to try to reboot the Windows server under any circumstances and always restart just the service? Just for cases when PRTG is co-hosted on a server that runs some other applications that just need to stay online..
PRTG usually only restarts the server when an update was initiated through Auto-Update as this requires a reboot in most cases. This can be avoided by disabling automatic installation of new versions and running the update when you want it to.
Also in PRTG Administration Tool you find options for scheduled restarts of PRTG's services or the whole server.
Besides those two reasons PRTG would not initiate an unattended reboot of the server.