health | A simple and flexible health check library for Go

def _check_health(self):
    while True:
      if self._check_health_thread_should_stop.is_set():
        return
      for job in self._cluster_spec.jobs:
        for task_id in range(self._cluster_spec.num_tasks(job)):
          peer = "/job:{}/repl  

def _start_check_health_thread(self):
    # Use a dummy all-reduce as a barrier to wait for all workers to be up,
    # otherwise the check health may fail immediately.

    # Use array_ops.identity to create the dummy tensor so that we have a new
    

def check_collective_ops_peer_health(self, task, timeout_in_ms):
    """Check collective peer health.

    This probes each task to see if they're still alive. Note that restarted
    tasks are considered a different one, and they're considered not h