clusters start: make the command idempotent when cluster is already active

[sridhar-3009]

Problem

databricks clusters start <cluster-id> fails with an error when the cluster is already running:

Error: INVALID_STATE: Cluster <id> is in unexpected state Running

The Databricks documentation states:

If the cluster is not currently in a TERMINATED state, nothing will happen.

The CLI was propagating the API's INVALID_STATE response as a hard error instead of treating it as a no-op.

Fix

Added a startIdempotentOverride in cmd/workspace/clusters/start_idempotent.go that wraps the generated command's RunE. Before calling the API, it fetches the current cluster state. If the cluster is already in any non-TERMINATED state (RUNNING, RESIZING, PENDING, etc.), it logs a message and returns nil, matching the documented behavior.

The original RunE is still called for TERMINATED clusters and for cases where no cluster ID is available before the picker runs.

Tests

Added unit tests covering:

• RUNNING cluster → no-op, no error
• RESIZING cluster → no-op, no error
• TERMINATED cluster → proceeds to call start normally
• PENDING cluster → no-op, no error

Fixes #1372

[github-actions]

An authorized user can trigger integration tests manually by following the instructions below:

Trigger:
go/deco-tests-run/cli

Inputs:

• PR number: 5776
• Commit SHA: 4118fc4326b1e63b63bfe7cc9505778bee5311c5

Checks will be approved automatically on success.

[github-actions]

Waiting for approval

Based on git history, these people are best suited to review:

• @pietern -- recent work in cmd/workspace/clusters/

Eligible reviewers: @andrewnester, @anton-107, @denik, @renaudhartert-db, @shreyas-goenka, @simonfaltum

_{Suggestions based on git history. See OWNERS for ownership rules.}