fix(server): send proxy errors as a2a errors

[jezekra1]

There are still cases when you will get http error code:

• calling provider that does not exist (http 404)
• no permissions to call provider (http 403)

Other errors are now correctly propagated through as a2a errors with status code 200 and error payload

The original motivation for this issue was to solve an invalid payload type for the client:
Expected response header Content-Type to contain 'text/event-stream', got 'text/html', however this is a a2a client issue, not possible to solve on the server side. It occurs when the agent card contains invalid transport, which we can do nothing about on the server.

I at least improved the error message, this is what's shown if your card has url "example.com":

❯ agentstack run chat hi
⠇ Extracting knowledge nectar...💥 RuntimeError: The agent is not reachable, please check that the agent card is configured properly.
Agent connection info:
  {
    "additionalInterfaces": null,
    "preferredTransport": "JSONRPC",
    "url": "https://example.com"
  }
Full Error:
  HTTP Error 400: Invalid SSE response or protocol error: Expected response header Content-Type to contain 'text/event-stream', got 'text/html'

💡 HINT: Are you having consistent problems? If so, try these troubleshooting steps: agentstack platform delete to remove the platform, and agentstack
platform start to recreate it.

Ref: #1497

• no docs needed

[gemini-code-assist]

Summary of Changes

Hello @jezekra1, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request significantly enhances the server's error handling for proxy requests, particularly those involving agent-to-agent (A2A) communication. The core change shifts the paradigm from returning various HTTP error codes for certain proxy failures to consistently encapsulating these issues within the A2A protocol's error payload, while maintaining a 200 HTTP status. This ensures a more uniform and predictable error reporting mechanism for clients, making it easier to parse and react to application-level errors. It also introduces a specific error type for invalid provider calls and includes a new test to validate this behavior.

Highlights

• Standardized Error Reporting: Proxy errors are now primarily reported as A2A errors with a 200 HTTP status code and an error payload, improving client-side error handling consistency.
• New Error Type: Introduced InvalidProviderCallError to specifically categorize issues related to calling providers that are in an invalid state (e.g., offline or error state).
• Delayed Agent Card Creation: The ProxyRequestHandler now uses an agent_card_factory to defer the creation of the AgentCard, allowing for better error propagation during the agent setup phase.
• Enhanced Provider State Checks: Added explicit checks in ensure_agent to raise InvalidProviderCallError when attempting to interact with providers that are offline or in an error state.
• Configuration Key Renaming: The JWT secret key environment variable in template.env has been renamed for clarity and consistency.
• E2E Test for Error Handling: A new end-to-end test confirms that offline providers correctly return A2A errors with the expected structure and status.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request refactors the error handling for proxy requests to ensure that server-side errors are propagated as A2A errors with an HTTP 200 status, rather than causing HTTP 500 errors. This is a good improvement for API consistency. The changes are well-structured, introducing a specific InvalidProviderCallError and using a factory pattern to delay agent resolution until it's within an exception handling block. A new end-to-end test validates the new behavior. I have a couple of minor suggestions to improve the clarity of error messages and test documentation.