Add HealthCheckRuntime context manager for shared boilerplate [1/2]

[gustcol]

Summary

Ref: #75

• Introduce HealthCheckRuntime, a @dataclass context manager that encapsulates the ~30 lines of repeated setup code every health check subcommand duplicates: logger initialization, GPU node ID detection, derived cluster resolution, ExitStack with TelemetryContext + OutputContext, and killswitch checking
• Reduces per-subcommand boilerplate from ~30 lines to ~5 lines (a with HealthCheckRuntime(...) as rt: block)
• Purely additive — existing checks continue to work unchanged; no migration in this PR

Stacked PR series: [1/2] Runtime helper → [2/2] Scaffold tool (depends on this PR)

Before (~30 lines per subcommand)

node = socket.gethostname()
logger, _ = init_logger(...)
try: gpu_node_id = gni_lib.get_gpu_node_id()
except: gpu_node_id = None; ...
derived_cluster = get_derived_cluster(...)
exit_code = ExitCode.UNKNOWN
msg = ""
with ExitStack() as s:
    s.enter_context(TelemetryContext(...))
    s.enter_context(OutputContext(...))
    ff = FeatureValueHealthChecksFeatures()
    if ff.get_healthchecksfeatures_disable_check_sensors():
        exit_code = ExitCode.OK; ...
    # actual check logic
    sys.exit(exit_code.value)

After (~5 lines per subcommand)

with HealthCheckRuntime(
    cluster=cluster, type=type, log_level=log_level,
    log_folder=log_folder, sink=sink, sink_opts=sink_opts,
    verbose_out=verbose_out,
    heterogeneous_cluster_v1=heterogeneous_cluster_v1,
    health_check_name=HealthCheckName.CHECK_SENSORS,
    killswitch_getter=lambda: FeatureValueHealthChecksFeatures()
        .get_healthchecksfeatures_disable_check_sensors(),
) as rt:
    # actual check logic using rt.logger, rt.node, etc.
    rt.finish(exit_code, msg)

Test plan

• nox -s tests -- gcm/tests/health_checks_tests/test_runtime.py — 6 tests covering initialization, killswitch behavior, finish(), context nesting, GPU node ID failure
• nox -s lint
• nox -s format
• nox -s typecheck

[github-actions]

CI Commands

The following CI workflows run automatically on every push and pull request:

Workflow	What it runs
GPU Cluster Monitoring Python CI	lint, tests, typecheck, format, deb build, pyoxidizer builds
Go packages CI	shelper tests, format, lint

---

The following commands can be used by maintainers to trigger additional tests that require access to secrets:

Command	Description	Requires approval?
/metaci tests	Runs Meta internal integration tests (pytest)	Yes — a maintainer must trigger the command and approve the deployment request
/metaci integration tests	Same as above (alias)	Yes

Note: Only repository maintainers (OWNER association) can trigger /metaci commands. After commenting the command, a maintainer must also navigate to the Actions tab and approve the deployment to the graph-api-access environment before the jobs will run. See the approval guidelines for what to approve or reject.