Pipeline not adding new fields

1. Describe your incident:

Messages processed by newly created pipeline no longer appear in any searches or streams.

2. Describe your environment:

  • OS Information: Ubuntu 22 LTS
  • Package Version: Graylog 5.0.5 single-node

3. What steps have you already taken to try and solve the problem?

  • Tested multiple messages in the simulator - all pass and show correct new fields.

  • Checked the Message Processor Configuration as per other forum posts:

  • Tried searching all streams / messages, cannot locate affected messages.

  • Disable the pipeline allows the messages to flow correctly, so is a pipeline related issue.

  • Have double checked all rules / connections in the pipeline.

  • Pipeline rule below:

 rule "GeoIP lookup"
   to_string($message.EventID) == "312"
    let parsed = regex("(?:[0-9]{1,3}\\.){3}[0-9]{1,3}",to_string($message.full_message));
    let geo = lookup("geoip-lookup", parsed["0"]);
    set_field("src_ip_geo_location", geo["coordinates"]);
    set_field("src_ip_geo_country", geo["country"].iso_code);
    set_field("src_ip_geo_city", geo["city"].names.en);

4. How can the community help?

I need help correctly configuring the pipeline to set the new fields with the information i need.

The rule above is double back-slashed, but seems to remove that when i post

Hey @dansta87,

Can you see messages passing through the pipeline/rule and not failing, is there anything within the logs that would point to the issue.

Perhaps make use of debug to print the value of parsed[0].

I’m curious if you have any Indexing or Processing failures on the ‘System / Overview’ page.


The pipeline rule doesn’t appear to be changing anything else to cause messages to disappear (such as timestamp). Its possible elasticsearch/OpenSearch is throwing index errors if its unable to store the data for a given field.

Thank you for the reply.
It appears i am indeed getting index errors:

Not sure where the 1000 fields is coming from as the above pipeline is the only thing I have active.

I appreciate the replies as this helped me look in the right direction. I’ve split my indexes and modified my NXLog for Windows Event Log to be more precise and not just pull everything. I’ve also reduced my index rotation from 90 days to 30 days.

My index errors have stopped and the pipeline appears to be working now with messages being correctly modified with the new fields.

Thanks all :grin:

1 Like

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.