Reduce integration test cluster capacity usage

[hectorcast-db]

Summary

Reduces integration-test cluster capacity by replacing SDK waiter coverage with deterministic HTTP fixtures and removing redundant live tests.

Why

The cluster timeout test created a real cluster solely to exercise client-side polling and cleanup, while the libraries test used a live cluster to validate server-side package installation. The direct command execution test also duplicated the higher-level command integration coverage. These tests consumed cluster capacity without adding corresponding live SDK coverage value.

What changed

Interface changes

None.

Behavioral changes

None. SDK runtime behavior is unchanged.

Internal changes

• Converted the cluster create timeout test to HTTP fixtures covering create, repeated PENDING polls, timeout, delete, and terminated cleanup.
• Removed TestAccCommandsDirectUsage; TestAccCommands retains live command coverage.
• Replaced TestAccLibraries with HTTP fixture coverage for install polling from PENDING to INSTALLED, followed by uninstall.
• Removed the now-unused live-cluster helper and documented the test-only change in NEXT_CHANGELOG.md.

How is this tested?

• make fmt
• go test ./internal -run '^(TestClustersCreateTimesOutAndCleansUp|TestLibrariesUpdateAndWait)$' -count=1

[github-actions]

If integration tests don't run automatically, an authorized user can run them manually by following the instructions below:

Trigger:
go/deco-tests-run/sdk-go

Inputs:

• PR number: 1844
• Commit SHA: 33aa7d13019ee7b1763370136d0098752cfd3584

Checks will be approved automatically on success.