no records logged which show that httpd.service is dead

[lucky]

There are no records logged which show that httpd.service is dead. The error is logged in /var/log/messages, this can make it difficult for the administrator to trace the error because /var/log/messages contains the general purpose error logs for overall system.

As keystone service runs inside httpd, the non-availability of httpd can make the entire openstack environment stop working.

To avoid this issue, there can be two possible solutions:
Log the error message in component level log. Here the error message should be logged in keystone.log so that monitoring remains within keystone component. In this case, changes should be done in keystone client end.
High level monitoring mechanism such as email containing the error message can also be sent to admin so that monitoring can be made more convenient.

[Lance Bragstad (lbragstad)]

This bug report seems circular. Keystone relies on an http implementation to run, but this report expects the keystone service to report status of the http service when it isn't running?

For for first solution, keystone won't be able to log anything if the service running it is having issues. At that point, things are outside of keystone's control.

For the second solution, are you expecting keystone to send the monitoring email?

[Launchpad Janitor (janitor)]

[Expired for OpenStack Identity (keystone) because there has been no activity for 60 days.]