> +1 for pushing this to the public list soon (also for not exposing your grafana instance to the world) I have obfuscated the credentials and I am forwarding this to the ceph-devel. > As Lenz indicated, we have a set of quite functional dashboards and I have been collecting feedback from a variety of users on how to improve these. > My plan was to propose these dashboards for the Ceph project. Then two weeks ago Grafana v5 was released and it has a number of features that we > should incorporate. So since then I have been busy setting things up for Grafana 5 and am in the process of re-rolling the dashboards for that (with the > afore mentioned feedback). > I'd heavily recommend checking out the last grafana! We are also thinking about using grafana v5 (we have done some tests), we are little worried that it is not fully backwards compatible with grafana 4.x though. > As for collaboration: I'm all for it. Its probably a good idea to set this off with a meeting of interested parties. If everyone is down with that, I'm happy to > organise such a thing. I suppose we can set this up. All the interested parties should be in the CC of this e-mail thread. > Iiuc the whole service map implementation is not quite far enough yet. I.e. we will have to have a (rados) gateway specific exporter for now. >From what I have heard today, there have been some changes in master in this area and the rgw data should now be flowing to the ceph-mgr although ceph-mgr does not expose it, yet. What rgw-specific exporter did you have in mind? Do you want to use an existing or write a new exporter? Regards, Boris ---------- Forwarded message ---------- From: Jan Fajerski <Jan.Fajerski@xxxxxxxx> Date: Mon, Mar 12, 2018 at 8:32 PM Subject: Re: Grafana dashboards To: Boris Ranto <branto@xxxxxxxxxx> Cc: Lenz Grimmer <lgrimmer@xxxxxxxx>, lenz@xxxxxxxxxxx, rdias@xxxxxxxx, rimarques@xxxxxxxx, vtheile@xxxxxxxx, sebastian.wagner@xxxxxxxx, Paul Cuzner <pcuzner@xxxxxxxxxx>, Zack Cerza <zack@xxxxxxxxxx>, Kai Wagner <kwagner@xxxxxxxx>, Sebastian Krah <skrah@xxxxxxxx>, Lars Marowsky-Bree <lmb@xxxxxxxx>, Ju Lim <julim@xxxxxxxxxx>, Jesus Ernesto Puerta Treceno <epuertat@xxxxxxxxxx> On Mon, Mar 12, 2018 at 08:12:31PM +0100, Boris Ranto wrote: > > CC'ing a couple more people that might be interested in this topic. > > On Mon, Mar 12, 2018 at 3:19 PM, Boris Ranto <branto@xxxxxxxxxx> wrote: >> >> On Mon, Mar 12, 2018 at 2:37 PM, Lenz Grimmer <lgrimmer@xxxxxxxx> wrote: >>> >>> Hi Boris, >>> >> >> Hi, >> >>> thanks for getting in touch, much appreciated! >>> >>> Replying from my work email address and adding Jan, Sebastian and Kai, >>> who have been involved in the creation of the Grafana Dashboards that we >>> created for integration into openATTIC (and ship as part of SUSE >>> Enterprise Storage). >>> >>> Going forward, I'd like to suggest to move this discussion to >>> ceph-devel, just to make sure we can pick up additional thoughts and >>> feedback on this implementation. >>> >> >> I suppose we can move the discussion there although I am not sure if >> we want to share the username/password for the grafana server to the >> world. +1 for pushing this to the public list soon (also for not exposing your grafana instance to the world) >> >>> On 03/12/2018 11:31 AM, Boris Ranto wrote: >>> >>>> we have been recently working on a couple of grafana dashboards for >>>> the prometheus exporter (and node_exporter) and thought you might be >>>> interested in taking a look. >>>> >>>> We have a running grafana instance with the dashboards running here >>>> >>>> ###DELETED### >>>> >>>> The username/password pair is ###DELETED###. Please note that these are >>>> still work in progress and we are still updating them as neccessary. >>>> >>>> We would like to collaborate and get a similar data coverage with the >>>> dashboard_v2 dashboards. Please do share any thoughts/comments with >>>> respect to these dashboards. >>> >>> >>> Thank you for sharing! >>> >>> Looks good upon the first glance, especially with the alerting >>> integration, even though a tad bit crowded in some places :) >>> >>> What Prometheus exporters do you use in the background to collect all >>> that data? >>> >> >> We use the ceph-mgr prometheus exporter, node_exporter on each node >> and Paul Cuzner recently wrote a prometheus exporter for iscsi gw >> which we also want to use. >> >>> Do you plan to add these dashboards to the Ceph git repo? >>> >> >> Yes, it would be great if we could add these or similar dashboards to >> the ceph git repo. That is one of the reasons I reached out to you so >> that we could get a similar coverage with these and yours dashboards. >> I am looking forward on hearing how far did you get with your >> dashboards. As Lenz indicated, we have a set of quite functional dashboards and I have been collecting feedback from a variety of users on how to improve these. My plan was to propose these dashboards for the Ceph project. Then two weeks ago Grafana v5 was released and it has a number of features that we should incorporate. So since then I have been busy setting things up for Grafana 5 and am in the process of re-rolling the dashboards for that (with the afore mentioned feedback). I'd heavily recommend checking out the last grafana! As for collaboration: I'm all for it. Its probably a good idea to set this off with a meeting of interested parties. If everyone is down with that, I'm happy to organise such a thing. >> >> >>> To my knowledge we were also planning on submitting the dashboards that >>> we created for Prometheus after some cleanup work. @Jan, would you mind >>> sharing an update on the current status of this? >>> >>> You can see some screen shots of our approach here (embedded into the >>> openATTIC UI): >>> >>> https://www.openattic.org/galleries/oa-3.x-screenshots/ >>> >>> The corresponding dashboard YAML files are somewhat hidden in this git repo: >>> >>> https://github.com/SUSE/DeepSea/tree/master/srv/salt/ceph/monitoring/grafana/files >>> >>> It would probably make sense to do a side-by-side comparison and review >>> what we can bring to the table to converge these two implementations and >>> collaborate on them going forward. From what I recall, we received some >>> very good feedback on our dashboards from some larger production >>> environments that we're still working in incorporating. Making sure the >>> dashboards and graphs scale on large deployments can sometimes be >>> challenging. >>> >>> I noticed that you don't have any RGW-related dashboard yet, so maybe >>> that's something we could contribute to as a first step? >>> >> >> We did have a rgw dashboard but we are running this on luminous >> cluster and there are no rgw stats exported by prometheus exporter, >> there. Do you have any idea whether this was fixed in master (or do >> you have any patch for that)? Iiuc the whole service map implementation is not quite far enough yet. I.e. we will have to have a gateway specific exporter for now. >> >> -boris >> >>> Thanks again for getting in touch - I look forward to joining forces >>> around this effort. >>> >>> Lenz >>> >>> -- >>> SUSE Linux GmbH - Maxfeldstr. 5 - 90409 Nuernberg (Germany) >>> GF:Felix Imendörffer,Jane Smithard,Graham Norton,HRB 21284 (AG Nürnberg) >>> >>> > -- Jan Fajerski Engineer Enterprise Storage SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton, HRB 21284 (AG Nürnberg) -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
