A little late response, I had some time off.

Thank you for updating the monitoring stack. I understand you can't always 
backport everything, but good to hear you're considering it!

> Op 29-04-2025 06:57 CEST schreef Nizamudeen A <n...@redhat.com>:
> 
> 
> Good point, we are still on v2 of prometheus 
> (https://github.com/ceph/ceph/blob/main/src/python-common/ceph/cephadm/images.py#L27)
>  (2.51.0) and I see v3 (https://github.com/prometheus/prometheus/releases) is 
> there so a bump is needed there. We will take care of it along with the other 
> dependent ones.
> 
> As for loki and promtail, we are moving away from promtail and going to use 
> alloy (https://github.com/ceph/ceph/pull/62960), since promtail is officially 
> deprecated by grafana. Once the migration is done we will bump loki and test 
> them together.
> 
> I am not sure if we will backport the alloy thing to squid (could break 
> stuffs), but we can say more when we test it.
> 
> And as for grafana it'll be 11.6.0 which is what we will use: 
> https://github.com/ceph/ceph/pull/62827
> 
> Regards,
> Nizam
> 
> 
> On Fri, Apr 25, 2025 at 8:37 PM Sake Ceph <c...@paulusma.eu> wrote:
> > This thread is only about Grafana, but what about the other services like 
> > Prometheus, Loki, Promtail, alertmanager or node-exporter?
> >  
> >  > Op 19-04-2025 15:18 CEST schreef Daniel Brown 
> > <daniel.h.brown@thermify.cloud>:
> >  > 
> >  > 
> >  > FWIW - Grafana 12 is going to be GA soon. 11.6 or later is probably 
> > where you want to target for “stable” releases for the moment. 
> >  > 
> >  > 
> >  > --
> >  > Dan Brown
> >  > 
> >  > > On Apr 18, 2025, at 10:47, Sake Ceph <c...@paulusma.eu> wrote:
> >  > > 
> >  > > Really great you moved to the default grafana image and moving the 
> > customization to cephadm. This makes upgrading indeed easier.
> >  > > 
> >  > > The problem is that Grafana has a stranger release cycle and therefor 
> > not really compatible with release cycle of Ceph. This would mean one or 
> > two update(s) of the supported Grafana minor version. And maybe with every 
> > Ceph release, update to the latest bugfix release?
> >  > > I guess a bugfix version should always work with the same minor 
> > version of Grafana. The user can configure this via the provided tools for 
> > setting custom images.
> >  > > 
> >  > > We're only talking about Grafana, but this applicable for all the 
> > extra container images.
> >  > > 
> >  > > Thank you for all the hard work you're doing for the project! These 
> > are just my 2 cents :D
> >  > > 
> >  > > Kind regards,
> >  > > Sake
> >  > >> Op 17-04-2025 19:43 CEST schreef Nizamudeen A <n...@redhat.com>:
> >  > >> 
> >  > >> 
> >  > >> We deprecated [1] the use of quay.io/ceph/ (http://quay.io/ceph/) 
> > (http://quay.io/ceph/)ceph-grafana images from squid onwards due to the 
> > pain of building and maintaining those images by ourselves. The main reason
> >  > >> we were building a custom grafana image was to bundle the grafana 
> > dashboards and plugins inside the official grafana image.
> >  > >> 
> >  > >> But with https://github.com/ceph/ceph/pull/55615, we offloaded that 
> > to cephadm at runtime which removed the burden of building and maintaining 
> > it which is why
> >  > >> we are using the quay.io/ceph/ (http://quay.io/ceph/) 
> > (http://quay.io/ceph/)grafanaimage and its simply a mirror from the docker 
> > registry. And we are upgrading the grafana to 11.x in tentacle which is in 
> > progress
> >  > >> by @Afreen Misbah here: https://github.com/ceph/ceph/pull/62827.
> >  > >> 
> >  > >> We have two options for reef since we are still supporting reef 
> > actively.
> >  > >> 1. we can try and see if backporting that PR is possible, last time I 
> > couldn't because of cephadm binary differences but I could look again
> >  > >> 2. Just tag and continue to build a newer ceph-grafana:10.4.x image 
> > in quay. Once the image is there anyone can just go ahead and use it by 
> > updating the container image config.
> >  > >> 
> >  > >> Will check both of them next week after the holidays.
> >  > >> 
> >  > >> [1] https://docs.ceph.com/en/latest/releases/squid/#monitoring
> >  > >> 
> >  > >> Regards,
> >  > >> Nizam
> >  > >> 
> >  > >> 
> >  > >>> On Thu, Apr 17, 2025 at 7:43 PM Sake Ceph <c...@paulusma.eu> wrote:
> >  > >>> Squid is already on 10.4.0 and it looks like it using the default 
> > grafana container images (correct?).
> >  > >>> 
> > https://github.com/ceph/ceph/blob/squid/src/cephadm/cephadmlib/constants.py
> >  > >>> 
> >  > >>> I couldn't find any issues, but the company really doesn't like old 
> > software with know issues :)
> >  > >>> 
> >  > > _______________________________________________
> >  > > ceph-users mailing list -- ceph-users@ceph.io
> >  > > To unsubscribe send an email to ceph-users-le...@ceph.io
> >  > _______________________________________________
> >  > ceph-users mailing list -- ceph-users@ceph.io
> >  > To unsubscribe send an email to ceph-users-le...@ceph.io
> >  _______________________________________________
> >  ceph-users mailing list -- ceph-users@ceph.io
> >  To unsubscribe send an email to ceph-users-le...@ceph.io
> > 
> 
> 
> --
> 
> Nizamudeen A
> Sr. Software Engineer - IBM
> Partner Engineer
> 
> IBM and Red Hat Ceph Storage
> Red Hat (https://www.redhat.com/)
> 
> https://www.redhat.com/
_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io

Reply via email to