feat: distribute actors across cluster when restoring on startup

[mattjohnsonpint]

When starting the actor system we restore agents that were suspended. This shuffles the list retrieved from the database, which has the effect of distributing the actors across multiple nodes, if multiple nodes are being started simultaneously.

[Copilot]

Pull Request Overview

This PR introduces an improvement to distribute actor restoration across cluster nodes on startup by shuffling the list of restored agents.

• Introduces a new asynchronous restoreAgentActors function for querying and shuffling agents.
• Replaces the local agent map with an ActorExists check to prevent duplicate actor spawns.
• Updates the CHANGELOG with a new feature entry.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
runtime/actors/actorsystem.go	Refactors actor restoration by shuffling agents and handling spawning logic, replacing a local map check with an ActorExists verification.
CHANGELOG.md	Adds an entry for the new feature.

Comments suppressed due to low confidence (1)

runtime/actors/actorsystem.go:95

• [nitpick] With the removal of the localAgents map and usage of _actorSystem.ActorExists to check for running actors, please ensure that ActorExists covers all necessary cases to prevent duplicate actor spawns.

	actors := _actorSystem.Actors()