Skip to main content
The Whawit VS Code extension integrates with GitHub Copilot, providing a chat participant (@whawit) and language model tools that give Copilot direct access to your observability data.

​
Overview

With the Whawit Copilot integration, you can:
  • Ask @whawit about incidents and analyses in natural language
  • Use slash commands to fetch specific observability data
  • Let Copilot access Whawit tools automatically when analyzing code
  • Get AI-generated fix recommendations based on real incident data
This integration requires both the Whawit VS Code extension and GitHub Copilot Chat to be installed and active.

​
Prerequisites

  • VS Code with GitHub Copilot Chat extension
  • Whawit VS Code extension installed and logged in
  • A project selected in Whawit

​
Chat Participant

The @whawit chat participant appears in GitHub Copilot Chat when the Whawit extension is active.

​
Basic Usage

Type @whawit in Copilot Chat followed by your question:

​
Available Commands

The chat participant supports five slash commands for specific actions:

​
Command Reference

​
/fix

Fetches observability analysis data and formats it for implementing code fixes. Usage:
What it returns:
  • Summary of the issue
  • Root causes with confidence scores
  • Code issues with file locations and suggestions
  • Recommendations for fixing
After using /fix, Copilot has context about the issue and can help you implement the fix directly in your code.

​
/analyze

Fetches analysis context for understanding and investigating codebase behavior. Usage:
What it returns:
  • Analysis summary and severity
  • Detailed analysis text
  • Affected entities and services
  • Timeline of events

​
/explain

Provides a detailed explanation of the currently loaded analysis. Usage:
You must first load an analysis using /fix or /analyze before using /explain.
What it returns:
  • Full summary
  • Detailed analysis
  • Severity with justification
  • All root causes with evidence
  • Complete recommendations list

​
/incidents

Lists recent incidents for your selected project. Usage:
What it returns:
  • Incident summary with severity indicator
  • Current status (open, closed, resolved, etc.)
  • Affected service name
  • Creation timestamp

​
/history

Shows recent analysis history for your project. Usage:
What it returns:
  • Analysis titles and status
  • Timestamps
  • Severity levels
  • URNs for reference in other commands

​
Language Model Tools

Beyond the chat participant, Whawit registers five language model tools that Copilot can call automatically when relevant:

​
whawit_getAnalysis

Fetches full analysis data from Whawit including root causes, code issues, and recommendations. When Copilot uses it:
  • When you ask about recent issues or errors
  • When you reference a specific analysis by URN
  • When context about system problems would help answer your question
Input:
  • urn (optional): Specific analysis URN to fetch

​
whawit_getIncident

Fetches incident details including correlated analysis data. When Copilot uses it:
  • When you ask about incidents or alerts
  • When you mention severity levels (critical, high, etc.)
  • When investigating production issues
Input:
  • urn (optional): Specific incident URN
  • severity (optional): Filter by severity level

​
whawit_getCodeContext

Searches the project knowledge base for relevant architecture and technical context. When Copilot uses it:
  • When you ask how something works in the codebase
  • When you need architectural context
  • When looking for documentation about specific features
Input:
  • query (optional): Search query for finding relevant context

​
whawit_listAnalyses

Lists recent analyses from Whawit with URNs and summaries. When Copilot uses it:
  • When you ask for a list of recent issues
  • When you want to see analysis history
  • When choosing which analysis to investigate
Input:
  • limit (optional): Maximum number of results (default: 5)

​
whawit_getProjectInfo

Gets information about the current Whawit project including available contexts. When Copilot uses it:
  • When you ask about the project configuration
  • When checking what integrations are available
  • When verifying project setup

​
Example Workflows

​
Fixing an Incident

1

Check for incidents

Review the list and note any critical or high severity incidents.
2

Get fix context

This loads the most recent analysis with fix-oriented formatting.
3

Ask Copilot to implement

With the analysis context loaded, ask Copilot:
Copilot now has full context about the issue and can generate appropriate code changes.

​
Understanding System Architecture

1

Load project context

Copilot uses whawit_getCodeContext to search the knowledge base.
2

Get detailed analysis

This fetches relevant analysis data about the payment system.
3

Ask follow-up questions

Get a detailed breakdown of the loaded analysis.

​
Daily Incident Review

​
Follow-up Suggestions

After using @whawit commands, the chat participant provides contextual follow-up suggestions:
  • After loading an analysis: “Explain Analysis” and “Fix Issues” options appear
  • After any command: “View History” option for quick access to analysis list
Click these follow-up buttons or continue the conversation naturally.

​
Troubleshooting

  1. Ensure the Whawit extension is installed and active
  2. Check that you’re logged into Whawit (look for the status bar indicator)
  3. Restart VS Code to reload the extension
  4. Verify GitHub Copilot Chat extension is installed and working
  1. Ensure a project is selected in Whawit
  2. Verify the project has analysis history and/or incidents
  3. Check the Output panel (View > Output > Whawit) for errors
  4. Try running /history first to confirm connectivity
  1. Copilot decides when to use tools based on your prompt
  2. Be specific in your questions to trigger tool usage
  3. Mention “Whawit” or “observability” to hint at relevant tools
  4. Use explicit @whawit commands for guaranteed tool access

​
Tips for Best Results

Be specific with queries: Instead of “what’s wrong?”, try “what Redis errors occurred in the last hour?”
Reference analyses by URN: When you see a URN in /history output, use it directly: @whawit /fix urn:history:abc123
Combine with code context: After loading Whawit context, ask Copilot to look at specific files: “Now check src/services/redis.ts for the issue mentioned in the analysis”

​
Next Steps

VS Code Extension

Full documentation for the Whawit VS Code extension.

MCP Server

Connect other AI assistants to Whawit via Model Context Protocol.