[ProblemChild] Bug in Ingest Pipeline for blocklist_label String/Integer - #7577
[ProblemChild] Bug in Ingest Pipeline for blocklist_label String/Integer#7577MakoWish wants to merge 5 commits into
blocklist_label String/Integer#7577Conversation
…ub.com/MakoWish/integrations into bug_problemchild_pipeline_string_value
…ub.com/MakoWish/integrations into bug_problemchild_pipeline_string_value
|
@MakoWish - Could you confirm this has been tested? |
This has been tested. I was not getting any matches at all for the |
susan-shu-c
left a comment
There was a problem hiding this comment.
This makes sense to me re: testing it, but will confirm with Apoorva when she's back!
|
@MakoWish Here's the blog that has all the details of the package installation steps. Step 2 under the Getting Started section states the mappings that need to be added to your index, which includes setting I did, however, notice that we have linked to the wrong blog in the package docs, so if anything, we should update that link with the following so folks don't run into this in the future: https://www.elastic.co/security-labs/detecting-living-off-the-land-attacks-with-new-elastic-integration If you want to change this PR to update that link instead, that'd be great, otherwise we'll cover it in a follow-up PR. |
I have the mappings set already, but that unfortunately does not resolve the issue. The mappings ensure the value is indexed correctly, but the problem is within the "inference" ingest pipeline itself (before indexing). The pipeline creates a string value of You will see the resulting document with |
|
@MakoWish Ah I see what you're saying. An alternative to adding an additional processor to the pipeline could also be to change the if check in the blocklist script processor to compare with "0" instead, or do the conversion to int in Painless using |
Please see #7618 |
Up to you. All options would accomplish the same. I just happened to go with the |
|
Let's go with comparing against a string for simplicity. Here's the result of a simple test: Result: |
|
Also, can we consolidate all these bug /broken link fixes into a single PR resulting in a single patch version bump? Thanks! |
|
Consolidating into #7618 |
* Update docs with new blog link * Commit changes from code review Co-authored-by: Apoorva Joshi <30438249+ajosh0504@users.noreply.github.com> * Commit changes from code review Co-authored-by: Apoorva Joshi <30438249+ajosh0504@users.noreply.github.com> * Consolidate #7521 and #7577 into #7618 --------- Co-authored-by: Apoorva Joshi <30438249+ajosh0504@users.noreply.github.com>
* Update docs with new blog link * Commit changes from code review Co-authored-by: Apoorva Joshi <30438249+ajosh0504@users.noreply.github.com> * Commit changes from code review Co-authored-by: Apoorva Joshi <30438249+ajosh0504@users.noreply.github.com> * Consolidate elastic#7521 and elastic#7577 into elastic#7618 --------- Co-authored-by: Apoorva Joshi <30438249+ajosh0504@users.noreply.github.com>
Type of change
What does this PR do?
This PR adds a
convertprocessor to ensure theproblemchild.predictionvalue is an integer before comparing to an integer value.Checklist
changelog.ymlfile.manifest.ymlfile.Related issues
blocklist_labelString/Integer #7576