Agent0 generates pull requests for infrastructure-as-code configurations, instrumentation improvements, and code-level fixes identified during investigations. Agent0 creates the changes, writes a description of the issue, includes evidence from the investigation, and raises the PR in GitHub for review.

How It Works

Agent0 analyzes your request and investigation findings, then automatically:

Generates the necessary code or configuration changes
Creates a detailed PR description with context and evidence
Includes investigation findings, telemetry screenshots, and rationale
References related issues or incidents if applicable
Raises the PR in your GitHub repository for team review
Validates that changes follow established patterns and conventions

The PR is reviewable like any other code change, enabling your standard approval and testing workflows before merging.

Note

Creating pull requests requires the GitHub connector to be configured. See Connect to GitHub for setup instructions.

Generate Pull Requests

Request a pull request after creating infrastructure-as-code or identifying a code-level fix:

Example prompts:

"Raise this Terraform configuration as a pull request"
"Create a PR to add these resource attributes to the service"
"Generate a pull request to fix the missing trace correlation"
"Raise a PR to improve the OpenTelemetry configuration"
"Create a pull request with this dashboard IaC"
"Submit a PR to add the service.namespace attribute"

Tip

Specify the target repository and branch if you have multiple repositories configured. If you don't specify, Agent0 will use the default repository from your GitHub connector configuration.

Types of Pull Requests

Agent0 generates pull requests for several categories:

Infrastructure-as-code — Terraform configurations for dashboards and alerts
Kubernetes manifests — Perses dashboards and check rule custom resources
Instrumentation fixes — Missing resource attributes, span conventions, correlation IDs
OpenTelemetry configuration — Sampler settings, processor pipelines, exporter configurations
Service metadata — Service names, versions, environment labels
Code-level fixes — Issues identified during root cause analysis

Review the Pull Request Details

Before raising the PR, Agent0 shows you:

Files changed — The specific files and modifications included
Description — Context about why the change is needed
Investigation evidence — Telemetry screenshots, error logs, trace examples
Target repository — Where the PR will be created
Branch strategy — Base branch and proposed feature branch name

Example PR description generated by Agent0:

markdown

1234567891011121314151617181920
## Add service.namespace attribute to frontend service

### Problem
The frontend service is missing the `service.namespace` attribute, which
prevents proper correlation with Kubernetes resources and makes it harder
to filter traces by environment.

### Evidence
- Service appears as "frontend" without namespace qualification
- Unable to filter by namespace in service map
- Related services in the "production" namespace show proper correlation

### Changes
- Added `service.namespace=production` to OTEL_RESOURCE_ATTRIBUTES
- Updated deployment manifest with environment variable
- Follows semantic convention v1.24.0

### Testing
- Verified resource attributes appear in spans after restart
- Confirmed service map now shows namespace grouping

Refine Before Raising

If the PR needs adjustments before creation, iterate with Agent0:

"Change the target branch to develop"
"Add a link to the incident in the description"
"Include the error rate graph in the PR description"
"Split this into two PRs, one for each service"
"Add a comment explaining why this threshold was chosen"

Raise the Pull Request

Once satisfied with the changes, confirm with Agent0 to raise the PR:

"Raise this PR"
"Create the pull request"
"Submit this to GitHub"

Agent0 creates the PR and provides a direct link:

PR #1234: Add service.namespace attribute to frontend service https://github.com/your-org/your-repo/pull/1234

The PR follows your team's review process — assign reviewers, run CI checks, and merge when approved.

Instrumentation Pull Requests

Agent0 reviews your service's OpenTelemetry configuration against established patterns and identifies gaps:

Missing or incorrect service names
Incomplete resource attributes
Missing span conventions
Log and trace correlation identifiers not configured
Sampling configuration issues

Agent0 generates the instrumentation code or configuration to close these gaps and raises it as a pull request:

Common instrumentation improvements:

Adding service.name, service.namespace, service.version
Configuring deployment.environment attribute
Setting up trace-log correlation with trace_id and span_id
Fixing span naming to follow semantic conventions
Adding custom resource attributes for domain context

Create Pull Requests