IBM Support

QRadar: How to find non-Linux OS events getting into Linux Log sources

How To


The Linux® OS DSM for IBM® QRadar® is designed to parse Linux operating system events. Events from other applications installed on top of a Linux system might be incorrectly identified and parsed as Linux OS events. Depending on the average event rate, incorrect payloads can reduce parsing performance, accuracy of log analysis and correlation.

It is more common to see this behavior for log sources or applications for which no available DSM exists. Correctly tuning log sources to handle those events can improve parsing performance.


Improve overall parsing performance




Note:  Use the Log Source Management App to verify that you are using the latest DSM before performing any troubleshooting or making changes.

How to identify these events

To create the search:
  1. Log in to the QRadar UI.
  2. Click the Log Activity tab.
  3. To create the First filter.
    1. Click Add Filter.
    2. For Parameter: choose Log Source Type (Indexed) > For Operator: choose Equals > For Value: choose Linux OS.
    3. Click Add Filter.
  4. To create the Second filter.
    1. Click Add Filter.
    2. For Parameter: choose Stored For Performance > For Operator: choose Equals > For Value: add False.
    3. Click Add Filter.
  5. To Create the Third Filter.
    1. Click Add Filter.
    2. For Parameter: choose Category (Indexed) > For Operator: choose Equals > For Value: choose High Level Category unknown > choose Low Level Category Stored.
    3. Click Add Filter.

The last filter is the most important because events can be categorized as stored for two reasons, either there are performance degradation issues or they cannot be parsed due to the nature of the payload. The second filter, "Stored For Performance" is False, displays only events that need to be addressed. 

Log activity filters
Though these events might be coming from multiple sources, it is possible that only a few are reporting a high rate. 
To identify the top sources:
  1. Use the search created.
  2. From View select or add a time interval.
  3. From Display select Source IP
  4. Click the Gear to expand Top 10 Source IP Results by count.
    image 11497
  5. Change the Value to Graph from Count to Event Count (Sum)
    Note: The log sources getting the most traffic are at the top. If their event count is significantly greater that the other log sources, then focus on them as they cause the more impact on performance.
  6. Double-click the top Source IP to open events only from that source.
  7. Set the Display Filter to "Raw Events" in order to more easily see what type of payloads you are getting. The service name generating the events is usually after the hostname or IP address. Identifying the service name is useful as with that information you can tell if the events either:
    1. Belong to a different log source type for which a DSM exists.
    2. There is no DSM and a custom one needs to be created.
    3. They are unwanted events. 
  8. Repeat step #6 for other the other top log sources in the list.
Results: If the events are any of the supported events types as described in Linux DSM Page: Linux OS - IBM Documentation, then open them to inspect the payload further and confirm they are not being truncated. Events are truncated when they exceed the Maximum TCP payload size. 

Check the event parsing statistics

To identify in which Event Collector or Event Processor these events are being received:
  1. Use the Search created.
  2. Click Search > Edit Search.
  3. Scroll to Available Columns.
  4. Search for "Event Processor" and move it to the Group By: section.
  5. Search for "Event Collector ID" and move it to the Group By: section.
  6. Click Search. The results include the Event Processor name and the Event Collector ID.

    Event Collector IDs
  7. To easily map the Event Collector's ID with its name, go to the Log Source Management App, and in the left column the Event Collector names and IDs are displayed. In the image, eventcollector115 corresponds to the host with name qradar74-ec

    Target Event Collector
  8. Open an SSH session into the identified Event Collector and collect parsing statistics by using:
    /opt/qradar/support/ -p 7777 -b "com.q1labs.sem:application=ecs-ec.ecs-ec,type=filters,name=DSM,id=LinuxServer"

    The parsing statistics give an idea of how slow or fast is the parsing on the Linux DSM. Be careful as the output is different on different collectors. It is important to correctly identify the Event Collector. 
    Note: On a QRadar on Cloud deployment, you can do this only on the Data Gateways.
    /opt/qradar/support/ -p 7777 -b "com.q1labs.sem:application=ecs-ec.ecs-ec,type=filters,name=DSM,id=LinuxServer"
    AverageEventParseTimeInMilliSecs: 0.4618693719426812
    EventsReceived: 4,889,237
    EventsParsedCount: 4,892,563
    EventsNormalizedProperly: 1,360,424
    EventsUnrecognized: 3,519,800
The variable AverageEventParseTimeInMilliSecs varies between log source types. Lower parsing times mean more efficient parsing. In this case, 0.46 ms per event is a high value indicating that this DSM is only able to process events at about 2,173 EPS. In the example, the high proportion of Unrecognized Events to Events Normalized Properly indicates the high average parse time is probably an effect of seeing events that do not match this DSM.

How to fix these events

Do the unparsed events belong to a different log source type for which a DSM exists?
  1. If another log source is already configured with the correct type, then check the Parsing order. Refer to QRadar: The use of Parsing orders.
  2. If all events are from a different log source type, then change the type directly from the "Log Source Managment" app
What if there is not an existing DSM that matches the unparsed event data?
If there is no DSM, then a custom DSM needs to be created. Refer to Creating a custom log source type to parse events - IBM Documentation
What if the events are unwanted?
Contact your Linux system administrator to reconfigure your sources to prevent them from sending the unwanted events.
Do not create a routing rule to drop these events because routing rules require events to be parsed first in order to drop them.
Refer to the QRadar DSM guide on how to configure the Log Source for supported events.
What if the events are getting truncated, causing problems with event parsing?
Depending on the size of the incoming payloads, you might see better performance by increasing the maximum payload size.

Opening a case with support

If the issue seems to be different from what is described in this article or you need other assistance identifying the problem, you can open a case with support. To speed up the process, run these commands on the QRadar managed host seeing the performance issue and upload the generated files to the new case:

  • for i in {1..5};  do /opt/qradar/support/ p 7777 --full >> ecs-ec-threads_$(hostname -s).txt; done
    • The output file is ecs-ec-threads_*.txt
  • /opt/qradar/support/
    • This command runs for ~ 2 minutes and generates a file named CustomProperties-*.tar.gz
  • /opt/qradar/support/
    • This command generates a file named ecs-mbeans.tgz

When opening a case with IBM QRadar Support, attach to the output files from the previous three commands and a get_logs report for the Console and Managed hosts.

Document Location


[{"Type":"MASTER","Line of Business":{"code":"LOB24","label":"Security Software"},"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSBQAC","label":"IBM Security QRadar SIEM"},"ARM Category":[{"code":"a8m0z000000cwtiAAA","label":"Performance"}],"ARM Case Number":"","Platform":[{"code":"PF016","label":"Linux"}],"Version":"7.4.0"}]

Document Information

Modified date:
20 September 2021