perf(clusters): avoid eager node loading in list operations

[srnbckr]

Description

Every cluster operation that called list_clusters() , including exls clusters list and the _resolve_cluster_id_callback that runs before every command with a cluster argument, eagerly loaded nodes for every cluster. For 5 clusters, exls clusters list made ~16 API calls when it only needed 1.

• Introduce ClusterSummary base model (without nodes) and make Cluster extend it, following the list-projection vs detail-projection pattern already used in the CLI app
• ClusterRepository.list() now returns List[ClusterSummary], skipping _load_cluster_nodes() entirely and therefore eliminates 2-3 API calls per cluster (get_cluster_nodes, list_nodes, get_cluster_resources)
• ClusterRepository.get() still returns full Cluster with nodes, unchanged

Notes for Reviewers

For a setup with e.g. 4 clusters, this reduces the time for listing the clusters with exls from >5s to ~1s.

Before:

________________________________________________________
Executed in    5.30 secs    fish           external
   usr time    1.01 secs    0.42 millis    1.01 secs
   sys time    0.09 secs    1.08 millis    0.09 secs

After:

____________________________
Executed in    1.26 secs      fish           external
   usr time  846.58 millis    0.00 micros  846.58 millis
   sys time   64.51 millis  685.00 micros   63.82 millis

[Copilot]

Pull request overview

This PR optimizes cluster list operations by introducing a lazy-loading pattern to avoid eagerly loading cluster nodes. For a setup with 4 clusters, this reduces the CLI execution time from ~5s to ~1s by eliminating 2-3 API calls per cluster during list operations.

Changes:

• Introduced ClusterSummary base model (without nodes) that Cluster extends, implementing the list-projection vs detail-projection pattern
• Updated ClusterRepository.list() to return List[ClusterSummary] instead of List[Cluster], eliminating calls to _load_cluster_nodes()
• Updated all call sites (CLI apps, service layer, workspace provider) to use ClusterSummary for list operations while get() operations still return full Cluster with nodes

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
exls/clusters/core/domain.py	Added ClusterSummary base model and made Cluster extend it with nodes field
exls/clusters/core/ports/repository.py	Updated list() return type to ClusterSummary
exls/clusters/adapters/adapter.py	Modified list() to create ClusterSummary objects without loading nodes
exls/clusters/core/service.py	Updated list_clusters() return type and _validate_cluster_status() parameter type
exls/clusters/app.py	Updated type annotations for cluster list and resolution operations
exls/services/app.py	Updated type annotation for cluster resolution in service commands
exls/workspaces/adapters/provider/clusters.py	Updated type annotation for cluster listing in workspace provider
tests/unit/clusters/test_clusters_service.py	Added ClusterSummary test fixture and updated list_clusters test expectations

---

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