You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Remove event.original removal processors from ingest pipelines #10072
It seems that most of the checked integrations have a removal processor for event.original (except if it contains the preserve tag) in their pipelines.
In the other way, the .fleet-final-pipeline also has this processor and runs after all integration pipelines (incl. the @Custom pipelines) have finished. If the integration has this removal, the @Custom pipeline cannot parse the original event anymore as it is not present in the event anymore.
I would suggest removing the removal processors from all integration ingest pipelines and forward the removal functionality into the fleet-final-pipeline where it belongs. In that case @Custom pipelines may parse the original event and can extract additional data or correct malformed parsing in the generic integrations (e.g. catalina.out in the Apache Tomcat integration):
I think this is a good idea. I will occasional install a logs@custom pipeline that keeps the event.original when there is a pipeline failure so that I can debug it.
This work was partially completed through several pull requests. To finish this out, I think we should script the changes. We can review edits to each team's integrations separately to optimize the process. Basically the script needs to delete the remove processor for the event.original field and ensure that the integration requires at least >=8.11.0 (where the Fleet final pipeline changes were introduced).
@Alphayeeeet, this issue was auto-closed by github in #13518. I'll reopen it.
Be mindful of using the word close next to the issue in the PR description, as github will interpret this as it should close the issue when the PR merges.
I usually write Related: #10072 to avoid this situation if the parent issue has a number of sub-issues/PRs under it.
Activity
andrewkroh commentedon Jun 6, 2024
I think this is a good idea. I will occasional install a
logs@custompipeline that keeps theevent.originalwhen there is a pipeline failure so that I can debug it.And in order for this to work, I currently have to manually drop that
removeprocessor from the managed pipelines.Alphayeeeet commentedon Jun 20, 2024
If that change would be approved, I maybe start working on this after my current PR has been merged.
Remove event.original removal processors (integrations 1Password to A…
Remove event.original removal processors (integrations B* to C*) (#10897
Remove event.original removal processors (integrations aws to azure) (#…
andrewkroh commentedon Dec 9, 2024
Relates #7636
andrewkroh commentedon Dec 9, 2024
This work was partially completed through several pull requests. To finish this out, I think we should script the changes. We can review edits to each team's integrations separately to optimize the process. Basically the script needs to delete the
removeprocessor for theevent.originalfield and ensure that the integration requires at least >=8.11.0 (where the Fleet final pipeline changes were introduced).Remove event.original removal processors (integrations 1Password to A…
Remove event.original removal processors (integrations B* to C*) (ela…
Remove event.original removal processors (integrations aws to azure) (e…
Remove event.original removal processors (integrations 1Password to A…
Remove event.original removal processors (integrations B* to C*) (ela…
Remove event.original removal processors (integrations aws to azure) (e…
Alphayeeeet commentedon Apr 12, 2025
Opened additional PR's to finally fulfil this issue. Please review accordingly
Removed event.original processors from network device integrations (#…
Alphayeeeet commentedon Jul 3, 2025
@taylor-swanson This issue is not complete until the following PR has been merged: #13518
taylor-swanson commentedon Jul 7, 2025
@Alphayeeeet, this issue was auto-closed by github in #13518. I'll reopen it.
Be mindful of using the word
closenext to the issue in the PR description, as github will interpret this as it should close the issue when the PR merges.I usually write
Related: #10072to avoid this situation if the parent issue has a number of sub-issues/PRs under it.Alphayeeeet commentedon Jul 7, 2025
@taylor-swanson Thank you for mentioning that. I must admit, I wasn't aware of that. I will keep it in mind for future PR's.