ADR-0044: Autonomous Provider Hot-Reload & Dynamic Capability Lifecycle v0.3.6

  • Title: Autonomous Provider Hot-Reload & Dynamic Capability Lifecycle v0.3.6
  • Status: Accepted
  • Date: 2026-11-12
  • Author: Community Core Team

Context

In long-running daemon environments, swarm operators need to update, reload, or deprecate community extensions without restarting the core kernel process. Abruptly tearing down a provider can corrupt in-flight transactions or leave abandoned goroutines. A coordinated hot-reload protocol with graceful draining and atomic registration replacement is required.

Decision

  1. Graceful Draining & Lifecycle Coordination: - Introduce ProviderLifecycleManager managing dynamic reload phases: Active, Draining, and Decommissioned. - When a reload or unload is triggered, new requests are rejected or routed to candidate versions while existing in-flight executions drain under a configurable timeout (default 10s).

  2. Atomic Swap & Dynamic Capability Update: - Provider registrations are swapped atomically under reader/writer synchronization without race conditions. - Capability unregistration cleans associated lifecycle hooks and topic subscriptions safely.

  3. Ontological Traceability & Quality Gate: - Traceable directly to parent Community Ecosystem Extensibility goals and active execution priority plans. - Complies fully with POL-DOC-001 character budgets and prohibition of internal cycle sprint narratives in public docs.

Consequences

  • Community extensions can be updated in real-time with zero host restarts and zero in-flight data loss.
  • In-flight execution draining guarantees transactional consistency.
  • The open-core extension architecture provides high availability and operational excellence.