Add persistent retry queue for failed telemetry events to Roo Code Cloud

[dleffel]

What specific problem does this solve?

Problem

Currently, the Roo Code Extension sends telemetry to Roo Code Cloud using a "fire and forget" approach. When telemetry events fail to send (due to network issues, server downtime, or other connectivity problems), they are simply lost with no retry mechanism.

This results in:

• Data loss: Important usage analytics and error reports never reach the cloud
• Incomplete metrics: Missing telemetry creates gaps in understanding user behavior and system performance
• Poor offline experience: Users working offline lose all telemetry data

Proposed Solution

Implement a persistent cloud related telemetry event queue system that maintains a FIFO (First-In-First-Out) structure where events are only removed from the queue after receiving successful confirmation from the cloud service endpoint. The queue should automatically attempt to process pending events whenever a new telemetry event is added, but only if the queue processing is not already in progress. Do not implement connection status detection or automatic retry mechanisms - the queue should simply try to send events when new items are added and retain all events until they are successfully transmitted to the cloud service.