Adds ML supervised model DGA package#2352
Conversation
3fd4939 to
109d650
Compare
packages/ml_dga/changelog.yml
Outdated
There was a problem hiding this comment.
Should the type here be enhancement? Or could this be something that's descriptive of the model, for example dga?
There was a problem hiding this comment.
Updated to 'ml-dga' so it's more descriptive in 7e5dafe824acfedaf66d9e2f4680a41095753495
packages/ml_dga/docs/README.md
Outdated
There was a problem hiding this comment.
Is this supposed to link to something? The link right after it? Usually I see square brackets in docs and think it's supposed to point somewhere
There was a problem hiding this comment.
yes, this links to https://www.elastic.co/guide/en/security/7.13/detection-engine-overview.html - we don't have anything about those models in the ML guide, so this is likely the only place that mentions them.
There was a problem hiding this comment.
I think there might be an extra space in the doc maybe? When I view the file, it's not hyperlinked if that's what's intended
There was a problem hiding this comment.
Should the version of the docs be updated to the 8.0 page - https://www.elastic.co/guide/en/security/8.0/detection-engine-overview.html?
There was a problem hiding this comment.
I'm assuming this is leftover from the ProblemChild PR?
There was a problem hiding this comment.
Yep - good catch - fixed in 7e5dafe824acfedaf66d9e2f4680a41095753495
There was a problem hiding this comment.
Nitpicky admittedly but the user sees these descriptions right? Misspelling in "comand"
There was a problem hiding this comment.
Same with 'algorithm' in the DGA parenthetical
There was a problem hiding this comment.
Thanks! Updated in 7e5dafe824acfedaf66d9e2f4680a41095753495
|
@alvarezmelissa87 How do tests look for this? If you could provide a quick screen recording of how this looks/is supposed to look, that'd be cool :) Might help clarify maybe any assumptions made in testing that a user may not know (I'm happy to try to replicate that if you need it at all, lmk) Update: I'm seeing this issue with demos for ProblemChild: https://github.com/elastic/ml-team/issues/628 with work being done on DGA. I'll follow this to see DGA demos, please let me know if that's not the case where I should be looking once that's done |
7e5dafe to
bf502fb
Compare
packages/ml_dga/manifest.yml
Outdated
| title: "ML DGA" | ||
| version: 0.0.1 | ||
| license: platinum | ||
| description: "ML solution package to detect domain generation algorithm (DGA) activity in your network data." |
|
cc @lcawl for package/model description text in the readme and such. |
packages/ml_dga/docs/README.md
Outdated
There was a problem hiding this comment.
Should the version of the docs be updated to the 8.0 page - https://www.elastic.co/guide/en/security/8.0/detection-engine-overview.html?
packages/dga/elasticsearch/ingest_pipeline/ml_dga_ingest_pipeline.yml
Outdated
Show resolved
Hide resolved
mtojek
left a comment
There was a problem hiding this comment.
The file packages/dga/elasticsearch/ml_model/dga_1611725_2.0.json is relatively big (~44MB).
Could you please tell us how many package revisions do you plan to add? Is it only this single one or do you plan to iterate on it in the next PRs? For example, 3 iterations will add an extra ~130MB to the Docker image, which is ~10% of all packages. I'm not sure if we have the capacity for this.
@mtojek only a single revision for this initial DGA package is planned. We are working with the docs team for example, to make sure the README does not require any further PRs, and with the security team to ensure that the package contains all the required detection rules before promoting the package to production. Looking further ahead, I do not anticipate that the assets (ML model and anomaly detection job, security detection rules) will need frequent updates. The last update to the model for example was released in the security detection rules repo in May 2021. Is there any update on the plans to redesign parts of the package registry to provide better support for large files - as discussed in elastic/package-registry#670 and https://github.com/elastic/obs-dc-team/issues/570 for example? |
|
Thanks for the confirmation, Pete.
We're working on this now, but considering the team capacity, it will take a few months to fully switch. |
lcawl
left a comment
There was a problem hiding this comment.
One suggestion, otherwise text LGTM
ba6d93d to
3fd49c9
Compare
|
|
||
| In **Machine Learning > Anomaly Detection**, when you create a job, you should see an option to `Use preconfigured jobs` with a card for Domain Generated Algorithm (DGA) detection. When you select the card, you will see a pre-configured anomaly detection job that you can enable depending on what makes the most sense for your environment. Note this job is only useful for indices that have been enriched by the ingest pipeline. | ||
|
|
||
| ### (Optional) Enable Security rules |
There was a problem hiding this comment.
We generally try to stick to sentence-case
| ### (Optional) Enable Security rules | |
| ### (Optional) Enable security rules |
|
This has been updated and is ready for a final look when you get a chance 🙏 cc @mtojek |

What does this PR do?
Adds the ML supervised model package for DGA model.
Package includes:
Checklist
changelog.ymlfile.Author's Checklist
How to test this PR locally
Related issues
Screenshots