monitoring and reporting storage connectivity

[Dafna Ron]

I think it would be nice if we had a real time health check for remotely configured storage (for cinder and glance) reported in horizon

[Ladislav Smola (lsmola)]

Do you have any suggestions how to measure the health? We could add a new pollster into Ceilometer for this.

[Dafna Ron (dron-3)]

actually, two things...
1. I think that this really needs a bit of a discussion since live health check, in my opinion, should be done considering different deployments and should be done really, really well to avoid performance issues and false reporting.
2. I think this should not depend on Ceilometer but be a default in Horizon itself.

I will alert some people to this RFE and for ideas so we can perhaps come up with a good design :)

[Matthias Runge (mrunge)]

In general, "liveness" is something, one would hide from the normal cloud user. I agree, it's useful for a privileged user (somehow).

Isn't this something, one would catch with nagios (or else?) Then we would also have reporting, if storage space get's below a certain point.

I would strictly disagree, to define a technical method here (on the horizon side), e.g. alert, if a storage has not been available the last 5 secs, 5 mins, .... (to be done in nagios)

[Liz Blanchard (lblanchard)]

Hi All,

I'm not sure how applicable these "Improvement to Horizon Overview page" wireframes might be to the idea behind this bug, but I wanted to post a link to the discussion thread here incase anyone has any thoughts or feedback. There is probably some networking information I'm missing here:
http://ask-openstackux.rhcloud.com/question/59/improvements-to-horizon-overview/

Thanks,
Liz

[Akihiro Motoki (amotoki)]

To achieve this, we need back-end service support. horizon itself does not provide a monitoring or healthcheck mechanism.
The scope of what need to be done in horizon is unclear, so I would like to mark this as Incomplete.

[Launchpad Janitor (janitor)]

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