Skip to main content
After CauseFlow identifies the root cause of an incident, it proposes specific remediation actions. Remediations are always suggestions — nothing executes automatically. Every proposed action requires explicit approval before CauseFlow takes any step.
Only admins and owners can approve remediations. Operators can view remediation proposals and track execution, but cannot approve or reject them.

How remediations work

When an investigation completes, CauseFlow generates a remediation plan based on the root cause and agent findings. The plan breaks down into individual steps — each step is a discrete action with a defined type, description, and expected outcome. You review the steps, approve or reject the plan, and then CauseFlow executes the approved steps in sequence.

Types of remediations

CauseFlow can propose several types of remediation steps depending on what the investigation found:
TypeDescription
Restart serviceRestarts a specific service or container in your infrastructure
Rollback deploymentReverts to a previous deployment version
Scale resourcesAdjusts the resource allocation for a service (for example, increasing instance count)
Create pull requestOpens a pull request in your GitHub repository with a proposed code fix
Custom scriptExecutes a custom remediation script defined in your integration configuration
Each step in the plan shows its type, a plain-language description of the action, and its current status.

Approval workflow

Remediations move through the following status flow:
StatusMeaning
proposedCauseFlow has suggested this action — no execution yet
approvedAn admin or owner approved the action — execution will begin shortly
executingCauseFlow is actively running the remediation step
completedThe step finished successfully
failedThe step encountered an error during execution
rejectedAn admin or owner rejected the action — it will not be executed

Approving a remediation

1

Open the remediation

Navigate to the incident or analysis that generated the remediation. Click the Remediations tab to see the proposed plan.
2

Review each step

Read the description of each step carefully. For infrastructure actions, confirm the target service and expected behavior. For code changes, review the diff before approving.
3

Approve or reject

Click Approve to allow execution, or Reject to decline. If you reject a step, you’ll be prompted to provide a reason — this helps improve future proposals and gives your team context in the audit trail.

Rejecting a remediation

When you reject a remediation step, enter a brief reason in the rejection dialog. For example: “This service restart would cause a 2-minute outage during peak hours — schedule for after midnight.” The reason is recorded in the audit trail and visible to your team.

Tracking execution

Once approved, remediation steps execute in order. The detail view updates in real time as each step runs:
  • A progress indicator shows which step is currently executing.
  • Each completed step shows its output — command output, API responses, or a link to the created pull request.
  • Failed steps show the error message and stop execution. You can review the failure, adjust your environment or configuration, and retry.

Pull request remediations

When the code analysis agent identifies a code-level fix, CauseFlow can create a pull request in your connected GitHub repository. The pull request includes:
  • A description of the root cause
  • The proposed code change
  • A reference back to the CauseFlow incident
The pull request appears in your repository like any other PR. Review it using your normal code review process — request changes, add comments, and merge when ready.
Pull request remediations require a GitHub integration. Go to Dashboard > Integrations to connect your repository if you haven’t already.