- Newest
- Most votes
- Most comments
Based on the error message you're seeing with your Bedrock AgentCore implementation using HealthOmics MCP via Docker, it appears you're encountering an internal server error when trying to perform basic queries like listing all runs.
This type of error ("An internal error occurred while processing the request") is typically a generic error that doesn't provide specific details about what went wrong. Here are some troubleshooting steps you can take:
- Check your container configuration:
- Ensure your MCP agent is deployed as a containerized application with host set to 0.0.0.0
- Verify you're using port 8000 for MCP protocol communication
- Confirm your container is using the ARM64 platform, which is required for compatibility with AWS Bedrock AgentCore runtime environment
- Test your container locally:
- Run your container locally and send test requests to verify functionality
- For HTTP protocol testing, you can use curl commands to the /invocations endpoint
- For MCP protocol, ensure your server properly implements stateless streamable-http and supports session management
- Check permissions and quotas:
- Ensure your Bedrock Agent has the necessary permissions to invoke any required services
- Verify you're not hitting any rate limits or quotas for Bedrock API calls
- Review logs:
- Check CloudWatch logs for more detailed error information
- Enable more verbose logging if possible to get better insight into the error
Since you mentioned this works on EC2 but not through the Bedrock API endpoint, it's possible there's a configuration difference between your environments or a specific requirement for the Bedrock integration that isn't being met.
If these steps don't resolve the issue, you might need to open a support case with AWS to get more specific information about the internal error you're encountering.
Sources
Understanding the AgentCore Runtime service contract - Amazon Bedrock AgentCore
Troubleshoot AgentCore Runtime - Amazon Bedrock AgentCore
Bedrock Agent hidden error | AWS re:Post
Bedrock agent "Internal Server Exception" | AWS re:Post
The message An internal error occurred while processing the request. suggests that the issue might have been in the Bedrock service. Is this an error you are seeing consistently or sporadically?
Relevant content
- asked 5 months ago
- asked 9 days ago

Can you say more about how the MCP is setup and secured, and how you have configured your agent to connect to it (and which Agent)?