the aws-ovn-serial job is failing at such a high rate that we are getting
alerts in the #ovnk-tiger-ci channel so much that it's noise is really getting
in the way of normal slack conversations.

this PR reduces the threshold from 0.6 to 0.1.

once this job is more healthy, and probably after the endpoint slices bug
is fixed, we should tweak the alert threshold back up to help us keep
the job stable.