Add a task in errorreports that reports all the errors to our telemetry server

[akolson]

Overview

Add a task in errorreports app that will report all the errors to our telemetry server

Description and outcomes

• As of now most of the errors are traced and are stored in ErrorReports model, so we shall be able to report those errors to our telemetry server.
• Create a task ping_error_report() inkolibri/core/errorreports/tasks/ that will be using ErrorReports.get_unsent_errors() to query unsent errors from ErrorReports and then make a post request to our telemetry API (to be implemented yet) with the errors. The JSON request structure will be:
  • For Eg:

[
  {
    "error_from": "frontend",
    "error_message": "custom message",
    "traceback": "custom traceback",
    "first_occurred": "2024-06-25T11:20:09.074Z",
    "last_occurred": "2024-06-25T11:20:09.074Z",
    "no_of_errors": 1
  },
  {
    "error_from": "backend",
    "error_message": "custom message",
    "traceback": "custom traceback",
    "first_occurred": "2024-06-25T11:20:09.074Z",
    "last_occurred": "2024-06-25T11:20:09.074Z",
    "no_of_errors": 1
  }
]

• After the task is created, enqueue it when our current ping_back mechanism has been successfully executed. So the enqueue can be done in a try/catch block within a try/catch block in kolibri.core.analytics.tasks._ping(). So the basic workflow will remain, if ping_back is successful, we enqueue errorr eport task, if error report fails we do nothing, so when ping_back is again enqueued and successful, the error-report task is enqueued again.
• After posting the errors, mark each posted error as sent(set sent of ErrorReport as true) using mark_as_sent() method

Acceptance Criteria

• ping_error_report() task created in errorreports/tasks.py.
• Error are marked as sent once reported
• The task is properly enqueued in kolibri.core.analytics.tasks._ping() when ping_once is successful
• Tests are written to ensure:
  • When the ping_back mechanism is successful, it enqueues ping_error_report()
  • When the task is enqueued, it filters unsent errors and makes a proper 200 request to the telemetry API

Assumptions and Dependencies

• As the telemetry API is not implemented, TDD should be followed.

Scope

• Creating the task and registering it
• Enqueuing the task in _ping()
  Out-of-Scope:
• Implementing telemetry API to receive the post request

Accessibility Requirements

NA

Resources

• Task Architecture
• Registering Tasks

[akolson]

@thesujai

[thesujai]

Ready to go

[akolson]

Closed by #12357