Fix pipeline reset consistency across backend state, simulation replay, and dashboard metrics

[alka104]

Summary

Improves pipeline reset consistency across backend and frontend components.

Changes

• Added reset helpers for Redis metrics, admin stats, in-memory threat engine metrics, and simulation state.
• Updated reset workflow to clear all backend state consistently.
• Preserved audit logs while continuing to record reset actions.
• Replaced destructive Kafka topic deletion with producer flush.
• Recreated Elasticsearch indices after reset.
• Added frontend handling for pipeline_reset events to clear dashboard counters, event logs, and globe state.
• Added race-condition protection during reset.

Result

Pipeline resets now clear alerts, metrics, simulation state, and dashboard visualizations consistently while preserving audit history and Kafka topics.