On Fri, Dec 8, 2017 at 1:38 PM, Viet Nguyen <vnguyen(a)redhat.com> wrote:
Clarification: I'm not suggesting an Assemble-Your-Own solution
whereby
the users independently install Prometheus then Hawkular Services.
Rather what I'm proposing is somewhat analogous to Fedora model. We want
to be the community leader in Prometheus on OpenShift
No we don't. This is not the goal of Hawkular Services and has never been
our goal.
Hawkular Services is about being able to monitor and manage middleware
servers. Thats our purpose.
We are a consumer of Prometheus and using it as our metrics solution. In
the past this was done using Hawkular Metrics. And in the future we might
switch to some other metrics solution.
We are currently using Prometheus as a tool. That is it.
There are other teams working on the Prometheus experience in OpenShift,
and we will be working with them. We will be reusing as much as possible
their containers, scripts and configurations. Its these teams whose goal is
the community leadership of running Prometheus on OpenShift, not us.
and on top of that our expertise in Middleware monitoring brings
values to
Middleware users by providing our Prometheus++ solution (core Prometheus +
Middleware monitoring enhancement).
>If someone wants p8s for something other than middleware
>monitoring, then they will have to use a different p8s pod.
Let me flip that inside out if I may. If someone wants to contribute to
P8S other than Middleware monitoring, then they will have to go to a
different project.
Yes of course, they will go to the OpenShift team working directly with
Prometheus. Just as we are.
With that I feel we may be missing a huge opportunity here. A large
community of P8s users will go elsewhere to contribute. We'll be on our
own island.
I'm bringing up the sidecar approach because naturally it imposes
loose-coupling between P8s and HS even when the two are in the same pod.
Viet
----- Original Message -----
From: "Matthew Wringe" <mwringe(a)redhat.com>
To: "Discussions around Hawkular development" <
hawkular-dev(a)lists.jboss.org>
Sent: Thursday, December 7, 2017 6:09:34 AM
Subject: Re: [Hawkular-dev] OpenShift Deployment
On Wed, Dec 6, 2017 at 3:17 PM, Viet Nguyen < vnguyen(a)redhat.com > wrote:
>Hi all,
>TL;DR: Technically it's ALL-IN-ONE POD but make HawkularService an option
aka "sidecar" container
I believe the 'sidecar' container concept is more used for when you have a
main container and then a small container to add in some functionality (eg
auth proxy, metric agent, etc)
In our case it would just be a pod with two containers which are tightly
coupled.
>Would our team also look at providing Prometheus (without HS) as
>the defacto choice for OpenShift?
For Hawkular Services, we will have our own Prometheus which is private to
our needs. Someone will not be able to use our pod and optionally only use
Prometheus from it.
>What I'm proposing is still technically an ALL-IN-ONE pod option.
> However, instead of looking at (Prometheus + HS) as a monolithic
>solution we can position HS as an enhancement to the plain vanilla
>Prometheus. This add-on sidecar[1] approach can satisfy both
>Middleware users and non-middleware community users who may not
>necessarily need HawkularServices. Let's say I want to use library >X and
X only comes with X+Y (which will cost me CPU and RAM
>resources) I may be less inclined to use the library.
We are not entertaining this idea and conceptually its closer to having
Hawkular Services as the main container and p8s as the side car container.
If someone wants middleware monitoring, they have to use our pod with
Hawkular Services and p8s. Its important that we control how our own p8s
instance works and to prevent someone from modifying it to their own
purposes (it would be too difficult to handle all the different scenarios
here).
If someone wants p8s for something other than middleware monitoring, then
they will have to use a different p8s pod.
[1] more on "sidecar" containers
http://blog.kubernetes.io/2015/06/the-distributed-
system-toolkit-patterns.html
Viet
----- Original Message -----
From: "Matthew Wringe" < mwringe(a)redhat.com >
To: "Discussions around Hawkular development" <
hawkular-dev(a)lists.jboss.org >
Sent: Monday, November 27, 2017 7:38:50 AM
Subject: [Hawkular-dev] OpenShift Deployment
With the changes that are now going to include Prometheus, how do we want
to deploy this in OpenShift?
We can have a few options:
ALL-IN-ONE CONTAINER
We put both Hawkular Services and Prometheus in the same container.
Pros:
- easy to deploy in plain docker (but this doesn't appear to be a usecase
we are targetting anyways)
- shares the same network connection (even localhost) and ip address (eg
but both services are on the different ports).
- Does't require any special wiring of components.
- Can share the same volume mount
- version of components can't get out of sync.
Cons:
- workflow doesn't work nicely. Docker containers are meant to only run a
single application and running two can cause problems. Eg lifecycle events
would become tricky and require some hacks to get around things.
- can't independently deploy things
- can't reuse or share any existing Prometheus docker containers.
ALL-IN-ONE POD
Hawkular Services and Prometheus are in their own containers, but they are
both deployed within the same pod.
Pros:
- shares the same network connection.
- bound to the same machine (useful if sharing the same hostpath pv) and
don' need to worry about external network configurations (eg firewalls
between OpenShift nodes)
- pvs can be shared or separate.
- lifecycle events will work properly.
Cons:
- lifecycle hooks will mean that both containers will have to pass before
either one will enter the ready state. So if Prometheus is failing for some
reason, Hawkular Services will not be available under the service.
- cannot independently update one container. If we need to deploy a new
container we will need to bring down the whole pod.
- are stuck with a 1:1 ratio between Hawkular Services and Prometheus
SEPARATE PODS
Hawkular Services and Prometheus have their own separate pods.
Pros:
- can independently run components and each component has its own separate
lifecycle
- if in the future we want to cluster Hawkular Services. this will make it
a lot easier and will also allow for running an n:m ratio between Hawkular
Services and Prometheus
- probably the more 'correct' way to deploy things as we don't have a
strong requirement for Hawkular Services and Prometheus to run together.
Cons:
- more complex wiring. We will need to have extra services and routes
created to handle this. This mean more things running and more chances for
things to go wrong. Also more things to configure
- reusing a PV between Hawkular Services and Prometheus could be more
challenging (especially if we are using hostpath pvs). Updating the
Prometheus scrape endpoint may require a new component and container.
_______________________________________________
hawkular-dev mailing list
hawkular-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hawkular-dev