refactor(cluster): read status from snapshot in scheduler

[mkoura]

The scheduling loop in get_cluster_instance queried the status database repeatedly - several queries per cluster instance in every iteration, all while holding the global cluster lock.

Add StatusSnapshot to status_db that loads all status records with one query per table. Its read accessors mirror the module-level query functions, including the filter argument semantics, so the scheduler call sites change only from status_db.* to self.snap.*. Reads from the snapshot are plain list traversals, which shortens the time the global lock is held and makes scheduling decisions testable against an in-memory state.

Writes keep going directly to the database. Every write done by this process bumps a module-level generation counter and the snapshot refreshes itself when its generation is stale, so mid-iteration writes (e.g. mark cleanup) stay visible to subsequent reads, same as with the direct queries. Writes done by other processes cannot happen while the lock is held - the scheduler refreshes the snapshot explicitly right after acquiring the lock.

The _respin method intentionally keeps the direct database queries, as it runs outside the global lock.

[Copilot]

Pull request overview

This PR refactors scheduler-side status reads to reduce repeated SQLite queries while holding the global cluster lock by introducing an in-memory StatusSnapshot in status_db. The snapshot loads all relevant status tables with one query each and serves subsequent reads from in-memory lists, while still ensuring freshness via a per-process write generation counter and an explicit refresh after lock acquisition.

Changes:

• Add status_db.StatusSnapshot plus a module-level _write_generation counter to support auto-refresh on local writes and explicit refresh for cross-process writes.
• Switch cluster_getter.get_cluster_instance scheduler read paths from direct status_db.* queries to self.snap.* snapshot accessors (refreshing once per lock acquisition).
• Add framework_tests coverage to validate snapshot parity with module-level read functions and refresh behavior.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
framework_tests/test_status_db.py	Adds tests asserting StatusSnapshot accessor parity and refresh behavior.
cardano_node_tests/cluster_management/status_db.py	Introduces write generation tracking, shared filter matching, and the StatusSnapshot implementation.
cardano_node_tests/cluster_management/cluster_getter.py	Updates scheduler reads to use the snapshot and refreshes it immediately after acquiring the global lock.

---

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