GitHub Copilot for JetBrains Adds OpenTelemetry Configuration, Model Management, and MCP Server Support

On July 27, 2026, GitHub announced new configuration features for GitHub Copilot for JetBrains, centered on OpenTelemetry export settings for agent workflows and expanded model management — making it easier to meet organizational observability requirements and manage costs.

Details

• OpenTelemetry export configuration: Available under “Settings > Tools > GitHub Copilot > Chat,” letting teams configure OpenTelemetry export for agent workflows to match organizational observability requirements
• Token limit controls: Default maxInputToken and maxOutputToken limits can now be set for BYOK and custom endpoints, supporting cost control and governance
• Model enable/disable management: A new model-management control lets teams enable or disable all built-in Copilot models at once, adding flexibility
• MCP server and custom agent support: MCP servers and custom agents can now be used directly within agent flows, letting teams bring specialized tools, instructions, and workflows into the IDE consistently
• Other improvements: The release also includes Copilot CLI fork support, a new /rubber-duck command, improved chat and model-selection UX, and better MCP diagnostics for faster troubleshooting

How to try it

• Update the GitHub Copilot plugin to the latest version in your JetBrains IDE (e.g., IntelliJ IDEA)
• Configure OpenTelemetry export under “Settings > Tools > GitHub Copilot > Chat”
• BYOK token limits and model enable/disable settings can be configured individually by admins or developers according to organizational governance policy