Constant ALARM state means only one actions event is ever emitted

[Steve Baker]

If a watch rule goes into an ALARM state and stays in the ALARM state, only one actions event is ever emitted.

In the example where bringing up one new server isn't sufficient to reduce the load below the alarm threshold, no new servers are ever created.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/34158

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/34158
Committed: http://github.com/openstack/heat/commit/bdeb33afdc52f37c2270a72c2221b051024f3f3f
Submitter: Jenkins
Branch: master

commit bdeb33afdc52f37c2270a72c2221b051024f3f3f
Author: Steve Baker <email address hidden>
Date: Mon Jun 24 14:19:36 2013 +1200

    Emit alarm actions if a rule remains in ALARM state.

    It looks like the original intent of this edge-triggering was to
    prevent actions being emitted too frequently, causing overshoot of
    scaling resources.

    However this causes an issue when one scaling resource change is not
    enough to bring the alarm state under the threshold.

    Proper use of Period, EvaluationPeriods and Threshold should be sufficient
    to prevent scaling resource overshoot.

    Fixes bug: #1193970

    Change-Id: I5b38eb82cd1ad568fe4af77999e9d52f9eb99328