collect-logs: set connect timeout value

collect-logs: set connect timeout value

If a job fails and the ceph cluster doesn't respond then the ceph status
task in the collect-logs.yml playbook will wait around 50 mins before
timed out.

  "cmd": [
      "ceph",
      "--cluster",
      "ceph",
      "-s",
      "-f",
      "json"
  ],
  "delta": "0:50:00.197544"
  "end": "2020-01-16 19:45:06.932819"
  "rc": 1,
  "start": "2020-01-16 18:55:06.735275"

  STDERR: [errno 110] RADOS timed out (error connecting to the cluster)

Instead we can use the --connect-timeout paramater to fail after 10s.

Signed-off-by: Dimitri Savineau <dsavinea@redhat.com>