Skip to content
Open
Show file tree
Hide file tree
Changes from 2 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
68 changes: 59 additions & 9 deletions hugo/content/en/bits_ai/bits_remediation.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,26 +9,76 @@
site_support_id: bits_remediation
---

## Suggested code fixes from Bits Code
After Bits Investigation helps you identify a root cause, it can also help you take action as quickly as possible.
{{< callout url="https://www.datadoghq.com/product-preview/bits-remediation/" >}}
Bits Remediation includes capabilities in Preview. Click <strong>Request Access</strong> to join the Preview program.
{{< /callout >}}

Bits Investigation integrates with [Bits Code][2] to automatically generate code fixes. Bits Code connects to GitHub to create production-ready pull requests, iterates on fixes using CI logs and developer feedback, and uses multiple Datadog products to generate contextual fixes.
1. [Set up Bits Code][1]. Then, after Bits Investigation has determined a code-related root cause, you will automatically receive suggested code fixes.
1. Ask Bits Code to make any additional updates as needed, create a pull request for review in GitHub, and merge when ready to fix the underlying problem.
## Automate code fixes

{{< img src="bits_ai/bits_ai_sre_suggested_code_fix.png" alt="Flowchart showing Bits' investigation conclusion and a suggested code fix" style="width:100%;" >}}
After Bits helps you identify a root cause from an investigation, it can also help you take action as quickly as possible.

Check warning on line 18 in hugo/content/en/bits_ai/bits_remediation.md

View workflow job for this annotation

GitHub Actions / vale

Datadog.words_case_insensitive

Use '' instead of 'quickly'.

Bits Investigation integrates with [Bits Code][2] to generate code fixes. Bits connects to your source code provider to create, update, and iterate on production-ready pull requests, based on existing issues detected by Datadog.

Code fixes are automatically generated by default for investigations with code-related root causes. To manually generate code fixes at your own discretion instead, disable automatic code fix generation in [Settings][3].

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Style: "at your own discretion" is filler and can be dropped; the surrounding sentence already conveys the choice.

Suggested change
Code fixes are automatically generated by default for investigations with code-related root causes. To manually generate code fixes at your own discretion instead, disable automatic code fix generation in [Settings][3].
Code fixes are automatically generated by default for investigations with code-related root causes. To manually generate code fixes instead, disable automatic code fix generation in [Settings][3].

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'd also rephrase the beginning of this sentence to "By default, Bits Investigations automatically generates code fixes..." to make the sentence active rather than passive 🙂


1. [Set up Bits Code][1]. After Bits determines a code-related root cause, a suggested code fix is generated in next steps by default.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggestion: "in next steps" reads ambiguously — is Next Steps a UI element in the investigation? If so, capitalize it (and match on line 25) so users know where to look. If it's not a labeled section, consider rephrasing to "as a suggested next step".

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's a bit hard to tell what this list is for. It's helpful to be really clear about when information is purely conceptual (in which case, maybe this list would be to give users an idea of the sequence of events involved in using Bits Remediation) or procedural (in which case, maybe this is a list of action items for the user to accomplish something). This list is a little murky with that distinction, which can be confusing for users. I think if we made it clearer, it would be really helpful!

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Adding an intro to the list like so!

To start using code fixes:

  1. Set up Bits Code ...
  2. ...

1. If automatic code fixes are disabled, manually generate a code fix from next steps.
1. Chat with Bits within the code session to update the suggested code fix.
1. Create a pull request for review and merge when ready.

{{< img src="bits_ai/bits_remediation/suggested_code_fix.png" alt="Bits Investigation hypothesis tree showing a root cause conclusion with a suggested code fix and other next steps" style="width:100%;" >}}

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe we can add an intro sentence before this image to explain exactly what it's showing?


## Run triage actions

## Run triage actions via chat
From chat, you can trigger triage actions without leaving the investigation workflow.

Supported actions include:
- Sending Slack and Microsoft Teams messages
- Creating incidents in Datadog and PagerDuty
- Paging engineers via Datadog On-Call
- Paging engineers with Datadog On-Call
Comment thread
jesshsiao marked this conversation as resolved.
Outdated
- Creating cases in Datadog Case Management
- Opening Jira tickets

Bits Investigation automatically pulls relevant context from the investigation and your connected integrations to prefill messages, incident descriptions, and ticket metadata. This reduces manual effort, ensures consistency, and accelerates response time.
Bits automatically pulls relevant context from the investigation and your connected integrations to prefill messages, incident descriptions, and ticket metadata. This reduces manual effort, helps ensure consistency, and accelerates response time.

Check warning on line 42 in hugo/content/en/bits_ai/bits_remediation.md

View workflow job for this annotation

GitHub Actions / vale

Datadog.words_case_insensitive

Use 'helps' or 'helps ensure' instead of 'ensure'.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Style: Nice catch swapping "ensures" → "helps ensure" here — matches the Vale rule. No change needed.


## Take action on your infrastructure

Check warning on line 44 in hugo/content/en/bits_ai/bits_remediation.md

View workflow job for this annotation

GitHub Actions / vale

Datadog.words_case_insensitive

Use 'act on' or 'take action on' instead of 'action on'.

<div class="alert alert-danger">One-click actions are in Preview.</div>

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggestion: The top of the page uses the {{< callout >}} shortcode for the Preview notice, but this section (and the Guardrails section on line 63) uses a raw <div class="alert alert-danger">. Consider using a consistent pattern across the page — either the shortcode or the inline HTML — so the visual treatment of Preview callouts matches.


For infrastructure-related issues, Bits can recommend a remediation action, such as scaling a deployment, restarting a pod, or patching a resource.

- **Manual recommendations**: Copy the suggested command (for example, a `kubectl patch` command) and run it in your own CLI.
- **One-click actions (Preview)**: Click **Run** to let Bits execute the suggested remediation action directly from the investigation context.

{{< img src="bits_ai/bits_remediation/one_click_action.png" alt="A suggested remediation action with instructions to restart a deployment, a Run button, and a Verify Resolution button" style="width:100%;" >}}

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm wondering if users might find this confusing because the Run button we're talking about is greyed out? Is there an available example we can use here where that button is usable?

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The only available examples are in staging/org2, not demo org -- we didn't want anyone in demo org to be able to run these actions. I can work with engineering to make the button not grayed out in demo org but not be able to apply actions, but would it be possible to start with this screenshot?

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
{{< img src="bits_ai/bits_remediation/one_click_action.png" alt="A suggested remediation action with instructions to restart a deployment, a Run button, and a Verify Resolution button" style="width:100%;" >}}
{{< img src="bits_ai/bits_remediation/one_click_action.png" alt="A suggested remediation action with instructions to restart a deployment, and a Run button to restart it directly from the investigation" style="width:100%;" >}}

We don't mention the Verify Resolution button in the copy (maybe we should?), so it doesn't make sense to only mention it in the alt text. I think doing that might make a user on a screen reader worry that they missed something if they start to wonder what that button does but can't find answers in the docs.

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this is related to the "Validate that issues were resolved" section on the bottom!


Kubernetes actions are supported in Preview, with more integration actions in progress. See the [Action Catalog][4] for the full list of supported actions and how to enable them in Datadog.
Comment thread
jesshsiao marked this conversation as resolved.
Outdated

To run one-click Kubernetes actions, your organization needs:
- A [Private Action Runner][5] with network access to your Kubernetes cluster, paired with a [Connection][6] to the Kubernetes integration.
- A user role with permission to run actions and resolve the Kubernetes connection.

## Govern how Bits takes remediation action

<div class="alert alert-danger">Bits Guardrails are in Preview.</div>

Bits Guardrails let admins define which remediation actions Bits can take, where those actions apply, and who needs to approve them.

Guardrails require the `Guardrails Read` and `Guardrails Write` permissions. To create a guardrail:

1. **Choose the actions to target**: Select one or more available actions for an integration (for example, Kubernetes) for the guardrail to target.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we add a step that tells the user where to go in Datadog to access these options? That will help users who come directly to this section through search 🙂

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@jesshsiao Ah, sorry, I was talking about where the Bits Guardrail options are! But the location of the permissions is helpful too!

1. **Define guardrail scope**: Specify the environment, service, and resource tags the guardrail applies to.
1. **Set the enforcement level**: For the selected actions and scope, decide when Bits can take action.
- **Ask**: Requires user approval before Bits executes actions. Choose which teams, roles, or individuals can approve.
- **Deny**: Bits can recommend an action but cannot take it.

## Validate that issues are resolved

Bits can verify whether a remediation action was applied successfully, and whether the original issue was resolved.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The heading here makes it sound like there's a task for users to do to validate, but then it isn't clear how a user actually does that. Maybe there was more that was supposed to be here?


[1]: /bits_ai/bits_code/setup/
[2]: /bits_ai/bits_code
[3]: https://app.datadoghq.com/bits-ai/settings/source-code-integration
[4]: /actions/actions_catalog/
[5]: /actions/private_actions/
[6]: /actions/connections/
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading