loadgen: add `status` mode to the http command

[Copilot]

Loadgen has gotten pretty complex with all the different modes and configurations. The problem is that when misconfigured, it just reports a generic failure in metrics. This makes debugging downstream harder, as consumers don't know what the error is without accessing core logs directly.

This PR adds a new status mode to the loadgen HTTP command that reports the most recent loadgen error status, if any.

Changes

LoadGenerator enhancements:

• Added mLastErrorMessage member to track detailed error information
• Enhanced emitFailure() to accept and store descriptive error messages
• Added getLoadGenStatus() method that returns comprehensive JSON status
• Updated error paths throughout the code to provide specific error details

HTTP Command interface:

• Added "status" mode to the generateload HTTP command
• Returns JSON response with running state, failure state, error details, and basic statistics

Testing:

• Added unit tests for the new status functionality
• Added HTTP command handler tests

Usage

Check current loadgen status:

curl 'localhost:11626/generateload?mode=status'

Example responses:

When idle:

{
  "failed" : false,
  "running" : false,
  "total_submitted" : 0
}

When running:

{
  "duration_seconds" : 15,
  "failed" : false,
  "running" : true,
  "total_submitted" : 47
}

When failed with error details:

{
  "error" : "Failed to submit transaction after 10 tries, status: 2",
  "failed" : true,
  "running" : false,
  "total_submitted" : 23
}

This enables much easier debugging of loadgen issues without requiring direct access to stellar-core logs, making it particularly useful for monitoring systems and downstream consumers.

Fixes #4684.

---

💬 Share your feedback on Copilot coding agent for the chance to win a $200 gift card! Click here to start the survey.