Create periodic task that logs agent's state

[Eugene Nikanorov]

When analyzing logs it would be convenient to have an information about liveliness of agents that would create a short-time context of cluster health.

That also could be than used for some automated log-analysis tools.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/181633

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/181633
Committed: https://git.openstack.org/cgit/openstack/neutron/commit/?id=8ee51f253cac950787273ca117611b4660ed4523
Submitter: Jenkins
Branch: master

commit 8ee51f253cac950787273ca117611b4660ed4523
Author: Eugene Nikanorov <email address hidden>
Date: Thu May 7 01:06:09 2015 +0400

    Add periodic agents health check.

    In addition to periodic checks of L3 and DHCP agents
    add periodic checks of overall health of registered agents.
    Log total count of agents at debug level so it can be
    seen in logs of neutron-server.
    In case some agents found dead - log detailed info about them:
    Type of agent, last heartbeat, host.

    Change-Id: I5db81dad4e9e8325ad3fa3a3e6d5d2d0deb297dd
    Closes-Bug: #1453320

[Thierry Carrez (ttx)]

This issue was fixed in the openstack/neutron 8.0.0.0b1 development milestone.