chores: refactor replica state

[GrigoryPervakov]

Why

Some fields had unclean semantics; the replica Pod was fetched multiple times during reconciliation.

[Copilot]

Pull request overview

Refactors replica reconciliation/state tracking to avoid repeated Pod fetches and to clarify semantics around replica startup errors, while improving status condition messaging.

Changes:

• Split replica-owned resources (ReplicaResources) from observed runtime state (ReplicaState), and plumb Pod + startup-error details through reconcilers.
• Replace boolean “error” flags with structured startup error strings and surface them in ReplicaStartupSucceeded condition (with message truncation).
• Update ClickHouse/Keeper replica update-stage logic and adjust unit tests accordingly.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
internal/controller/statusmanager.go	Builds detailed replica startup-error condition messages and truncates to a safe max length.
internal/controller/resourcemanager.go	Uses observed Pod/startup error from replica state to decide when to delete stuck Pods, avoiding re-fetches.
internal/controller/replicastate.go	Introduces ReplicaResources vs ReplicaState, adds GetReplicaPod and PodStartupError.
internal/controller/keeper/sync.go	Switches Keeper reconciliation to use new replica state/resources and startup error reporting.
internal/controller/keeper/sync_test.go	Updates Keeper tests for the new replica state/resources model.
internal/controller/clickhouse/sync.go	Switches ClickHouse reconciliation to use new replica state/resources model and updated staging logic.
internal/controller/clickhouse/sync_test.go	Updates ClickHouse tests to match new UpdateStage behavior and resource struct changes.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.