Red Hat OpenShift OpenShift Coffee Break | Red Hat Livestreaming, 21 Sep 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: OpenShift Coffee Break: OpenShift GitOps

Description

Get your espresso ready for the EMEA OpenShift Coffee Break as it's a full GitOps hour!
We welcome our special guest Harriet Lawrence, PM for OpenShift GitOps, to talk about latest updates on the Argo CD community together with latest and greatest feature on OpenShift GitOps Operator.

Agenda:

- What is GitOps
- OpenShift GitOps and Argo CD
- What is a real life use-case for GitOps? How are customers using GitOps in their environments?
- New in v1.6
- Upcoming in v1.7

A

A

Welcome back to the openshift TV coffee break. Today we have a special episode with our special guest. We're gonna introduce our special guest in a moment before we go welcome to the open shift. Tv coffee practice is a weekly show. We have here in Canada near time zone uh where we talk all about Cloud native kubernetes and openshift. My name is Natalie winter product marketing manager for openshift and I'm. Here, with my co-host, Andrea.

B

Hi Andre, how are you hi very well today and welcome everybody I? Have it hello, hello,.

A

How are you so, let's introduce our special guest horrid Lawrence. Do you want to introduce yourself.

C

I'm, the uh PM for openshift get-offs.

A

That's great that's great, and for this you see my celebration t-shirt for uh welcome to the show. uh We have lots to talk about, I, know and I'm sure you all know. Ketops is a really trending topic and talking about gear UPS we have. We can also mention that, with there there was articlecon in these days um you can say I'm gonna share in the chat the link to for our present to argokon uh all right. First question: what is what is argument? What is what is Argo.

C

Oh so Agro CD is a open source project and it provides well in the name, continuous delivery um and ago Khan has been happening this week and it's a conference in the US all about Argo.

A

That's great, that's great I've, seen many pictures on Twitter uh from our colleagues uh being there. uh So it's very it's very cool conference base. It's a cncf element that is sponsoring also is a diamond sponsor and we have a keynote and a special um event. Today, the definition day detox uh happening in Sunnyvale, our office, I, didn't know we have this office uh fantastic, so folks, um I hope you had your coffee shot. Oh and I hope you. You are you're having one, because here we have to talk about uh here we go.

A

We have to talk about very interesting topic, all about get ups and I. Don't know, I heard it. If you have any think you want to present to start with.

C

I did I brought along some slides. If that's all right, it's got some visuals for folks, perfect. Let's get started all right, thanks cool, so welcome folks, um we can go on to the next slide, get to see my face: hello, I'm Harriet.

C

um So these are places you can find me um I'm on Twitter. You can shoot me an email if you've got questions about get Ops or openshift get-ups, uh and if you want to chat I'm in the cncf slack as well.

C

So this is a little overview of what I'm gonna share with you all today. So have a look at what is get Ops um openshift get Ops and Argo CD.

C

um How are people using get Ops in their environments and then we'll have a look at what we've released recently in our openshift get Ops product and what's coming up all right, so you probably all have heard about get Ops um I'm, not sure how many folks are like deeply familiar with it. So I thought we'd go over the get Ops principles to start with.

C

uh So these are the four core get Ops principles. They were defined by the get Ops working group back in 2021, so the first is that your system is described declaratively, so that means you've written down somewhere, usually in yaml. What your system should look like, um be that an application with access controls or how a cluster should be configured so.

C

Secondly, we want this desired state to be versioned and immutable, so we need our yaml files to be stored centrally and accessibly, and they create a single source of Truth for our system, so this usually comes about by using git, but any system that complies with this principle can be used.

C

Next, we want the changes that we approve to be applied automatically, so usually that approval process happens via a pull request to your config repo, um and once that merges we want those changes to roll out automatically.

C

So not only does it pull from git, but it pulls at regular intervals to check for changes which leads us to the last one. So this is the one that I feel is real key part of get Ops. You've got a controller that manages and monitors your repo, it pulls for changes and it detects drift between your desired State and the actual state of your system, and it can act on that drift. It could notify you of the discrepancy or it could do self-healing and automatically bring it back into sync.

C

So go to the next slide. This is openshift get up. So what? What is that compared to agree CD, so openshift get Ops is at its core uh kubernetes operator. It's a package of custom resource definitions and it's based on the open source project, Argo City.

C

So we provide the access to the ago CD UI to manage and monitor your deployments, um and you can use that through openshift, so openshift get Ops is our red hat official version um that we make sure to apply our standards of stability and security to, and you get a whole bunch of extra support when you use the openshift operator.

C

So what do you get in uh openshift get Ops in Argo, so it allows you to do multi-cluster, config management, so, um as I mentioned before it can manage cluster and application configuration, and it can do that for multiple clusters from one place. One of the most exciting recent features that's just become generally available. Is application sets so applications? That's automatically generate Argo CD applications so from a single yaml file. You can describe your application list.

C

All the Clusters that you'd, like it deployed to and Argo CD will take care of, creating them all at the same time, across as many clusters as you specified and there's lots more that you can do with application sets. um Creating and managing replicas is just one and we've also got automated, install and upgrade of Argo CD. So.

B

When you install get Ops.

C

Through I know what happened to our slides.

A

Oh, we lost a slide, but uh it was an interesting uh what to say. The I took this opportunity earlier to ask you more about application sets um because I I know it's a feature which is um arriving from from some version, but um I would like to ask you if application center, the foundation for deploying application to multiple clusters? What is um what what is the purpose of application sets.

C

Yeah so it's kind of like a way to template um your applications or your clusters, um so you define in one place what your applications should look like and say you want three replicas.

C

um You don't have to Define. You don't have to like write out the application three separate times you can do it in one place and the application set controller will take care of stamping those out for you and giving you multiple copies and managing them. For you.

B

Could also work in cases like um where you putting together a Dr solution, and you want to make sure that you are deploying on let's say both clusters: the DDR solution. The same version not only are the applications, but all the configurations, like all the resources that are required and make sure that they don't drift.

C

ah So we've got our slides back yeah, all right um cool, so we were talking about the automated installment upgrade. So um when you use operator Hub to install get Ops uh it'll do an instance of Argo CD for you and configure it for you. It's all you're ready to go um and you can choose to automate upgrades as well. So the operator.

B

C

Lifecycle manager, olm, I, don't often say these acronyms all out loud. So it's going to be a tongue twister for me this morning, uh will ensure that you're, always on the latest version um and that'll include any new features and any relevant security objects as well.

C

um We're not quite ready for the next slide. Oh sorry, thanks.

A

This backstage today is not working good. Do you need another coffee for us.

C

um Of course, so opinionated get UPS bootstrapping, so if you're not sure where to start um or if you just want to like play around with a non-trivial example, um something that can form the basis of your workflow, we provide a bootstrapping tool called cam.

C

um Cam is a CLI and it uses best practices when setting up environments and it can help you getting started with a directory structure. It'll create a technon pipeline for you with web hooks that connect back into agrocd and it's a really great way to familiarize yourself with githubs, if you're not ready to just make your own configurations, and it can give you a kind of jumping off point to customize your own setup.

C

Finally, we've got deployments and environments insights so get Ops is integrated with the openshift console, um you're probably used to working in the developer and admin consoles already, if you're an openshift user, and you can find your get UPS applications in there as well. You can see all your environments, you can filter by project, and you can see your deployment history in there right alongside everything else in the console all right. So now we're ready for the next slide.

C

Also, how are people actually using git Ops in their production environments, and the answer is a lot of different ways. So every customer that I speak to is doing things slightly differently, which I think really highlights the flexibility of get Ops and Argo CD.

C

Some folks are using it purely for application delivery, and this is this is probably the use case I see most often, and uh what I think a lot of folks assume from hearing the name: Argo CD, that's the cd part, The Continuous delivery, part of it.

C

um One setup that I saw was managing 18 000 applications across four different environments.

A

Wow 18 000 application. Oh that's Ah that's great. To mention.

C

A

Question after but that's great to.

C

Mention thanks yeah, so they had um four different environments like Dev test stage, prod something similar to that um they're running three different instances of Argo CD and managing access control with um ABAC and assigning um teams and users to what they needed uh so yeah. What was your question? Natalie yeah.

A

um How to scale uh I'll just get it out right. You say trade instances for 8 000 up, uh so they did they make any uh tuning, or is there any best best practices? You know how many, how many user, how many app anarco instances could support.

C

um So it really depends on sorry. Every answer I'm gonna give is, it depends yeah.

A

It's a wise answer.

C

um It depends on what you're doing with it uh if you've got lots and lots of namespaces and uh like not many applications or if you've got heaps of applications and not very many namespaces, um but you can. um You can adjust the um memory and CPUs that.

B

C

Have access to um to to give them a bit more oomph um and power more applications, uh so the folks running, 18 000 applications have done that. um uh But I'd say that that is a more rare case. We don't usually see people running that much stuff, but um we often get questions about like oh. How much can I handle like quite a lot so go wild foreign.

C

Other people are doing cluster config, um so they're uh using it to configure the Clusters themselves, um not just the applications, so um you can use like we were talking about application sets. You can declaratively Define your cluster config and then stamp out as many clusters as you need, which is a cool way of making sure that you know exactly how all your clusters are managed and what settings they've got and making sure that people aren't going in and manually adjusting what your clusters are doing.

C

um So one cool example of that I saw um a single instance of Argo on a hub cluster. Managing config for all the spokes, um I've also seen folks with an Argo CD per cluster, managing just the config for that cluster.

C

um So this kind of Blends in a little bit with the multi-cluster management piece, which is the the third dot point there. So you can totally use Argo CD to manage a single clusters config, but um more and more often, these.

B

C

Folks are running tens or hundreds of clusters, and that becomes a real pain to manage by hand. So you want to be doing it declaratively and have something managing that, for you.

A

Right, I'm um and for this approach already having a one Argo per cluster uh who's gonna manage the Argo, so you know you you're, reducing the complexity of having one Central article for all the cluster, but also adding a Mania complexity to manage those Argos. How do you see this uh use case.

C

You can use ago to manage Argo.

C

So you can kind of have the best of both worlds. You can have um one Argo on like a hub or um just another cluster that manages the rest of your Argos um or if, if you've got really stringent security requirements, and you need them, each of them to be really isolated from each other.

C

um That gets into the realm, where you're either going for um depth of security or ease of use and as you've got to make a trade-off there between, like what you've got capacity to manage and the security requirements that you have.

B

And yeah just a question from the Pina Gallery here, so um there has to be at least one component of Argo that runs and they call them the managed clusters in the sense that you said there is a controller that is is is checking if there is drift. So something is running on the on that cluster, even if you don't have the full Argos. That is that correct.

C

Yes, um so I didn't include a slide about like the how it all sits together. um That would have been great. But um yes, there there is uh the I believe it's the repo server or the application controller I'm like not overly familiar with the technical particulars.

B

But even even that one would be managed, like you said that configuration would be, could be managed by Argo as well. Yes, yeah, just in very interesting.

C

um And that kind of use case is becoming more popular. um I've found more often folks, aren't aware that you can use our go to manage Argo, but as soon as they discover that they're like oh, yes, we are going to start doing this and um they they get really into it and start using our go to manage everything.

C

um Cool, so for the multi-cluster management and multi-tenant application management piece. If we go to the next slide, we can look about like the different Arrangements of instances kind of like what we're talking about um and how it works with permissions and security, and so B3 are um different deployment strategies for Agro CD instances, um they're all totally reasonable ways of doing things. There's not one! That's better than another. One um I've seen people using each one of these successfully in production, or you know, development test using them successfully.

C

uh So whenever people ask about what is like the best way to set up ago, um like I said before, I always have to say, it depends uh and then ask questions so things like how uh your development team is structured. What access requirements do they have? What is your organization's threat model? Who who's going to be managing the instances and other people managing the instances different to who will be using the applications day-to-day so once you've got answers to questions like those and you've got a good understanding of your requirements.

C

Then you come and look in this slide, and one of them should make more sense than the others for your particular use case. So we can have a look at each one.

C

The one on the far left is the central Hub model, so you've got one Argo instance and that pushes configuration out into remote openshift or kubernetes clusters. So this is a nice way to get started. It provides a single place to administer, but it does require some elevated permissions to push out to the Clusters.

C

So if you've got more stringent security requirements, the middle option, the clusterscoped model- might suit you better.

C

So the Argo CD instance sits inside the same cluster that it's modifying and you can have as many clusters with their own Argo CD instances as you need um and having Argo inside the cluster means that it doesn't need to store any access credentials as it's.

B

C

Inside the cluster, so either of these ones are really great for multi-cluster management, but if you're looking for more fine-grained access control, then those two provide you could look at the application scope model on the far right. So this is more into the multi-tenant application management area as well.

C

um When you have multiple ago CD instances inside a single cluster, they can each manage a single application, um pulling config from different configuration repos and that lets you divide up your access permissions even further.

B

Just a quick question: could you have a combination of these approaches? So let's say, for example, we mentioned before somebody may want to use argot for cluster config and, uh and so they may have a central one for to manage and the classical figure of all the Clusters. But then they want to have application scoped for their application and have their own ago CD instances in each for each is that is that possible. Yeah.

C

Absolutely do whatever you like: okay, well,.

A

I have another question: the the example you mentioned before the 18 000 apps is that a Central app use case divided in three.

C

um I believe they've got clusterscoped so.

A

C

C

Yes, it was a while ago that I spoke to them um um yeah, so they had three different ago instances all managing different environments.

A

Oh and those are three big clusters or uh many clusters.

C

I believe it was three clusters. Oh.

A

Okay yeah, so it's a classical scope with three cluster. uh Okay got it cool, no just to um understand the two: how to scale it up right yeah. uh If I they made the 8 000 hubs, but if they have a 100 cluster, uh what would be the you know just to uh conversation or hear a chat just to understand what would be.

C

A

C

Totally but there's just so many different ways to use Argo City yeah, it's always great to hear about people, doing it in different ways and how.

A

C

A

And if is there anyone in the chat uh assisting that has experience with Vargo City? Please share your experience right in the chat. If you have any question, we will bring uh to the to the stage.

C

Foreign right so before we leave this slide, um speaking of like names based uh applications and things like that, there's a really great feature that our engineering team has been adding to Upstream um and I believe it merged just recently. So that's support for application uh CRS in any namespace, so not just in the control plane, namespace, um so I, this one's been discussed Upstream for quite some time, I believe um and it's finally in Argo CD proper. So it means that you can.

C

You can specify a list of namespaces that you want Argo CD to scan for applications instead of it only looking inside um the namespace that it's in, which is really cool.

A

So an application, an Argo app- could be cluster-wide, yeah interesting.

C

um So that that's available that will be available in the next version of agricd um and we'll be including that in the next version of openshift githubs as well.

A

And what will be the difference between uh namespace scope, application and a cluster-wide application? Is there any particular tag or any a new new CR? uh How does it works.

C

Oh, like um in the yaml yeah, um so you can specify a namespace when you're, creating your applications here um and I believe it's that you use one. It just depends which name space you um specify in that in that okay, um okay,.

B

C

Link off to some documentation, for you.

A

C

C

I will add that later, because otherwise I'm going to be like I'm.

A

Probably gonna add it to the pattern: yeah that that's interesting, I think uh the application I think application is becoming more and more a standard for defining an app in kubernetes. What do you think about.

C

Yeah um I, agree and I think with the rise in popularity of get ups, and um this kind of different way of doing things, um although, like I, say different way of doing things, it came out of devops. So it's not like completely foreign folks, um but I. Think it's a really great move that we're moving towards this way of doing things and I I hope to see it proliferate even further.

A

B

Right next slide.

C

Yes, thank you.

B

C

Yeah cool, so let's talk about what what we've given you recently, so we um ship the um most recent version of our Upstream Argo CD, along with each version of openshift, get Ops and we tend to track it fairly closely um and we release on a similar, Cadence, um so you're, usually getting the latest from upstream and within your openshift version of openshift get Ops so like we were talking about before application sets, it is gone GA, so um that was very exciting. Yay.

C

um So that was that went GA in version 1.6 um in the in the previous release, so in 1.5 we also included some improvements to application sets with the addition of two new generators.

C

um So one you can Auto discover open, pull requests in your config repo and another to merge parameters across generators for Selective overwriting, so I've linked off to a couple of code. Samples here um I believe we'll share the slides at some point. um Otherwise, I can drop those examples in the in the chat.

A

Yeah we can, we can take the link and share in the chat yeah.

C

I've got it here, I'll just drop it in. Oh I. Don't know how to how do I chat.

A

There should be a chat on your right, but if you don't see it, we can link it for you, those link like the the two generator. Oh you yeah. If you, if you break into the private chat, we can copy the publisher.

C

My first time using restream, it's it's a very pretty interface. It's lovely.

A

B

C

um uh Okay cool uh so uh as part of the Improvement for that we also added a new flag to sync policy. So previously you could only ignore parts of a manifest when you're doing the diff, uh but you can do that during sync as well.

C

So, for example, if you've got something that modifies a pod at runtime to have more memory, you don't want Argo to come back in during the think and say: oh, this is different I'm going to bring that back to what the Manifest says. You want it to keep that extra memory.

C

uh What else have we got notifications? So notifications is here as well. um We've made that available as Tech preview to begin with, um and that's why we we make sure that it meets our standards of stability and security. Before we put it generally available, uh so notifications means you can set up triggers based on events like Health, degraded or think successful, and then send out a notification in an email or via slack or one of the other available Services there's a whole bunch of them.

C

um Config management plugins- these are pretty exciting, um they've, had a really great update and they can now be configured via a sidecar to the repo server.

C

So this enables you to bring in the tooling that you need, even when it's not part of Argo CD core, so you can bring in specific versions of things like customize or you can extend Argo to use something like Sops for secret management.

C

um Encrypted calms so communication between Argo CD components and the redis cache is now using TLS encryption. So if you've got any sensitive data moving to and from the cache, it's got more protection during Transit. So that's a very nice addition, um Power and Z, so we've finally gone multi-arch. So this is the first step on our multi-arch journey. So as of version 1.6 openshift get UPS can run on IBM power.

B

C

Z, architectures and we've got more planned, which we'll talk about on the next slide, uh and finally, earlier this year we published a blog post on the guide to Secrets management with get Ops and kubernetes.

C

um This was written by our engineering, Tech lead on get Ops yarn Fisher and it walks through an overview and a discussion of the popular tools and different approaches to Secrets management that are currently available. So it's a really great read. Thanks for dropping the link in that's Harley yeah.

A

That's great to read was a very nice tactic, also just recommend everyone going for it, because this is a really hot topic right. uh Okay, everything is on git, but also password. uh How? How do you manage it and uh yeah and those approach are all valid, I think already correct if I Wrong, the the general approach for openshift would go into external Secrets like having this uh kind of uh uh layer where, where any third-party secret management system can be plugged in into the into the system,.

C

Yeah, so um that is suddenly a popular project. At the moment um we don't have like an official way of doing things yet um I'm, actually gonna talk about Secrets Management on the next side, if you'd like to.

B

Go to this side.

A

C

We can just skip to the last one there and talk about Secrets management first, uh so that is something that we get asked about all the time like. What should we do for Secrets management? You don't have anything built in uh so we we would love to add first class support for folks to use secret managers in Argo CD.

C

um We are looking at several different ways of doing this, external Secrets being one of them, um and we know we know that Secrets management is really important to our customers um and we want to make sure that we're making it as easy as possible for you to use your choice of Secrets manager. We don't want to be prescribing. This is the one that we support. You have to use this one.

C

um So yes, external Secrets, is something that we're looking at right now, um but we're also evaluating the other options.

A

C

Thanks- and there is some stuff around already so there is the ah what is it called I think it's the there's, a vault plug-in that you can use as a CMP and bring in Vault support.

C

um It's not like officially managed by us or anything, but it is available to you if you, if that is um what you need.

A

C

Cool so other stuff coming soon. um This is some stuff that we're working on at the moment upcoming in the next version or a little bit further down the road. So pretty much the further. You go down this list the uh further in the future. It is, um and they're kind of like wider that we're looking so none of it's like prescribed. We know exactly how we're going to do things um we love to hear from our customers and our users like. What do you need?

C

How should we do this um and we'd love to take that into consideration when we're figuring out how to build stuff all right? What have we got so first, ah we have image updater, so this is happening Upstream at the moment, and so when you're, when you're Building images in like just the eye pipeline, it you've got a new image and then it's a real pain to have to go and update your manifests with new tags or new digests for the image image.

C

Updater gives you the ability to auto update the Manifest once a new image is available. hmm That is going to be a really awesome addition. um So.

A

C

A

I'm wondering yeah that that what you mentioned is a popular approach right. You you're something update the image hash uh to the Manifest and then hargo will uh roll out, but uh I was wondering how Argo can understand that an image has been updated. Is there any partling to the image? uh How does it works.

C

um I don't know if I have a good answer to that and post that somewhere.

A

Yeah. Thank you uh just out of curiosity, but this feature is available in Tech preview in uh overshift gearbox. This means it's already available. Upstream.

C

But we're still working on Upstream so.

A

C

um Yeah, this is the like. Still in progress stuff, uh a lot of our Engineers on the get Ops Team do heaps of work upstream, and so um one of our team is working on implementing it, Upstream right now and it's yet to be merged. So once that makes it into Argo CD, then it will come down to overshift get UPS as Tech preview.

A

Cool uh it looks like from what you said there is that the app stream is really close to the downstream right. There's no much difference. Yeah.

C

Yeah, we don't maintain like a fork or anything um we just pull directly in from the latest version into our Downstream great.

C

So, let's have a side apply uh for patching resources, so auger CD, just recently added support for um so SSA came into kubernetes as like an improved version of cube. Control apply as server side apply, and this is a function where, if you're unfamiliar with it, if if multiple users are trying to manage the same field, um SSA will provide conflict management, and so you can, you can choose to force an overwrite and claim management of the field or you can.

C

When you see a conflict, you can say like actually, no I don't want to manage that field. I didn't realize it was managed by someone else and I don't actually want to edit that one- and you can see your claim.

C

um So this helps you when you're, trying to patch resources that are managed by something else, so that has merged into the Upstream um and it will be coming Downstream very soon and.

A

C

I have if, if you're unfamiliar with service I'd apply generally I, have a link for you as well to read a bit more about it. Just.

A

In there right.

C

General kubernetes documentation.

C

uh All right, cam, cam, bootstrapping tool, Improvement so I mentioned cam earlier uh helps you set up get up. Some pipelines provides you a folder structure and web Hooks and everything to test out. uh We are adding more options to the CLI and the API.

C

We're making sure that the folder structure is something that can set you up for what we consider best practice use of get UPS um if you are using cam or want to use Cam and you've got uh ideas for how to make it better or um that would suit your use case. Please do get in touch. We. We love to hear from folks um using the tool and and we're kind of re-evaluate re-evaluating how people are using cam right now. So it's a great chance to get in with your ideas.

A

That's interesting this because this tool um I put the link in the chat. Everyone can see the repo and join with a uh issue and proposal, but um this is cool just to mention. uh If I recall correctly, this is a generating a skeleton of a github's project where you have a pipeline and then you have the argosity bits. Is that correct, yeah.

C

Yeah, that's it interesting yeah! So it's like it's a command line tool and um you can specify options for how like what kind of stuff you want and point it in the right places um to the Clusters that you've got going and then it will spin up all the stuff that you need for a test environment.

B

C

It's pretty fun, um I used it just recently and I I thought it was delightful. It was a really nice way of just having a test environment um I needed it for a demo, and it was brilliant.

A

Fantastic, you see how technical RPMs in read that uh super technique.

A

So you are the first you you are the first user right. You can report back to the to the the software any any.

C

Experience to improve it yeah, that's great I feel like I've got like the The Secret inside track to our developers that can like report bugs directly to them, which is really nice um uh all right. What else have we got arm support so yay, um now that we are building multi-arch um arm is on the horizon.

C

um I know that open shift as a whole has got like a real push towards arm support, uh and we want to get on that bandwagon as well.

A

Well, if you think about the edge use case right, um multiple cluster, let's say one I'll go per cluster, one I'll go per device uh arm. Support is definitely a great I. Don't know if she already knows about Edge use case with openshift githubs. Just out of curiosity.

C

Yeah yeah we've got um quite a few customers with those kind of use cases that are um either already using git Ops for it or evaluating it and getting ready to roll out a whole heap of applications. Great.

C

Oh cool, what else have we got multi-tenancy and permissions guide, so I'm really excited about this one, um so ow developers and our documentation writers have been collaborating on a really big new piece of documentation um and it is a guide to how permissions and multi-tenancy work in openshift githubs and how to utilize those to create the access control that you need for your organization.

C

I, don't have a firm date of when that's going to be available, but keep your eyes peeled, because it's going to be awesome, it's really thorough and it's it's amazing.

B

C

A

B

Around that I've seen that you know even in in the list of features and capabilities, you've been talking about guides and uh and documentation, and you know have basically helped the users understand how to best utilize, Lego City in the key drops. Are there also or planned uh classes from Red Hat training that would allow people to uh to learn how to get the best out of it? How to the the do's and don'ts? If you wish.

C

C

Proper website um of video tutorials um I can't entirely. Would you be able to drop a link to those as well.

A

C

Thanks uh they're really awesome way to they're quite introductory, like uh get familiar with Git Ops uh and start using it. um I would love to add some more of those. uh So if you've got any suggestions, please let me know, and if anyone wants to drop suggestions in the chat or like tutorials or like video guides, that you would love to see, please let.

B

C

We'd love to hear about it.

C

Thank you all right lot. One Progressive delivery. So this one we get so many questions about and um is like a really exciting topic for folks at the moment. So we are, we are looking at lots of different ways to support um folks different deploy strategies. So if you want Canary.

B

C

Or blue green deploys that's two of the really popular ways of doing stuff um and as part of that, primarily we are evaluating agar rollout as a project. So even though they share the Argo name, Argo rollouts and Agro CD are separate code bases.

C

So unfortunately, we can't just like slip on support for I'll go roll out, um but our team is having a look at it right now.

C

um We even had a meeting about it just yesterday, so we will be providing some kind of an answer to Progressive deliveries for folks using openshift get Ops um and there is a reasonable possibility that it's going to be Argo rollouts.

C

uh So there will be something coming uh we are. We are working on it right now. I know we. We already have a lot of folks using Argo rollouts directly. You can you can use whatever you like. If you want to manually, install it and maintain it yourself, so I think we've already got about 1800 instances of rollouts running um in openshift get-ups, um so folks are using it and getting value out of it, and that's awesome.

A

Super super interesting progressing delivery has been one of the Hot Topic and we have some books talking about it. We will share after after in.

C

A

We have some book about uh published on developersfood.com progress. Delivery is also one of this topic. So uh very interesting. uh Did you put the secret as a last uh like that? You know best at least I. Don't know how to say in English. Here is best, at least that's.

C

The best one for last, um so yes, that is probably the furthest one away um but I I kind of stole everything I was going to say about it to talk about it at the start, um I was just going to say that we we're looking at different ways of doing this, and um so.

B

You did mention that you were looking at specific sequence managers. You mentioned one, but I didn't catch the name, because I wanted to look at it.

C

Oh um I mentioned that there is already a plug-in for Vault how she got Vault: okay, um yeah we're we're looking more broadly at things like uh external secrets that allow you to use whatever you're already using, instead of um specifying a specific one that you should use.

C

Cool, that's! That's all I've got for today.

A

uh I'm gonna I'm gonna share the link to external Secrets, since you mentioned it yeah, um uh and we have one question in the chat: uh there's Fabio asking uh that you would like to hear something about Argo, CD competition. So in this kit, Ops landscape, we know I was not the only player, but do can you mention uh others, and why Argo you think it's a good. Why opposing the top selected article as their primary.

C

Option yeah, so the main one that you'll probably hear about in the theme breath is Argo is flux, so that was created by weaveworks. um We've worked for the folks that are credited with coining. The term get Ops as well, um and um there are. There are a bunch of others that are playing in the same space, but I'll go and flux seem to be the ones that come up most often.

C

hmm um Why did we choose agrocity, uh so I'm actually fairly new to Red Hat I joined in February this year and our openshift get Ops product has been around since the start of last year, so I was not around for the the choice of choosing Argo CD um I do know that it is a delightful tool and we have a bunch of our. uh Our Engineers are Upstream contributors and maintainers.

C

um So perhaps it was that we were already invested in that community and we we thought it was looking very promising. I'm, not sure. That's.

A

B

If one goes and looks at the community uh site for ago CD, there is a huge list of adopters of article CD. My question is how many contributors, though Amy's redhead, one of the main contributors, the ones that has more commits.

C

um So I think everyone, so our our engineering team is about well how many people have. We got now be 10 to 15 people on our engineering team and they all commit Upstream.

B

But are there other organizations that.

C

B

Are playing a big part in it?.

C

um Yeah definitely um so we're just one of them. um There's also uh a few of the like Main folks who help the maintainers be uh be functional together from Intuit. They like help run the monthly calls and that kind of stuff and I know they have a big investment in agrocity as well, and they um have a lot of contributors on their side.

C

um The folks from Acuity and um I believe there's.

B

C

Folks in there there's people from all over um a nice collaborative community that we've got going there.

A

uh It's great and I can see from the argocon um it's an on-site event, but you can see the pads on socials and the and their website of the event. That's great content, I think the community is really growing uh and that that's great to see.

C

Yeah totally and I believe all the the recordings from agricon should be available online, um I'm, not sure when, but at some.

A

Great because we have some good talk from Red doctors uh over there, so we have also a lab. uh So if you had the chance to go, you still can go to the dev national github's day. Otherwise you can catch up with recording on YouTube I. Think after.

C

A while they publish on the cncf.

A

Youtube yeah cool um um about Community, Eric I have another question, so you mentioned you discussed the this feature in in Upstream from from uh from a while. You know you were discussing one of this feature, but if one want to join the arcosity community want to propose a feature.

C

A

C

Yeah, so there is the agriproge GitHub organization and that um you can go there and see all of the different repos that they've got where um Argo core lives and there's also other project Labs, which is where they kind of incubate ideas of things that could be part of Argo. One.

B

C

Kind of like slightly left of what they're currently doing um interesting things that you can add into your agrocity and I, would say, go and check out. What's going on up, there have a read through the issues and pull requests. They've got uh contributing guidelines there in the contributors markdown document.

C

um If you're more of a face-to-face person, there are Community meetups every month, um I believe they are on Thursdays.

C

B

C

Cadence I can't remember how often they happen. I assume they're monthly, yeah.

A

It's typical correctly, those are key, tops working group right, something um on slack cycle.

C

Correctly, that happens in the cncf slack um and there there's a bunch of really great channels in that slack as well about Argo and get Ops and um go there and see what other people are talking about. What questions they've got um and I chat with other folks who are already using those products.

A

Great and uh it's great to to know that that has put engineering efforts in the project. You say: 10 15 people working at uh on the project which is uh really cool so and uh it's really cool also uh where that is bringing uh uh business use case. No production use case back to app stream. No, uh is there anything missing, uh it went it goes upstream, and then you come back Downstream to the to the to the operator and talking about the operator.

A

um So the operator is what we deliver. Basically right, we um is it and you tell there's a the version, is close to argosity Upstream.

A

um The operator um is the one we support uh as with that, so the installation, via the operator. How does it work if you want to install operation githubs in your cluster.

C

Yeah, so um if you've got your uh openshift cluster running, you can go to operator Hub, you can search for get Ops and our openshift gearbox operator will appear for you. uh It has a really friendly, install you just click install and it will go and do everything for you. So you can choose what channel you want to install. You can pin to a specific version or you can just choose latest.

C

We actually did just recently deprecate the preview and stable channels and we've moved to latest. So it's called latest it's. It is stable. Please do not worry folks, um we are getting rid of stable and preview channels um and the latest most stable is available in the latest Channel and um if you stay in the latest, Channel you'll get the newest updates. You'll get the latest versions. All the latest patches, all that kind of stuff.

A

That Scream 2 here and since you mentioned patch, uh we Reddit, is also responsible on the security patch. You know on the when something happens, uh so if there's anything that need to be patient, the operator can automatically patch the the software. Yes.

C

Yes, it can, and um so because we are a vendor of agrocity, we get told about any um any problems that pop up as part of the Embargo period. So we will have a patch ready to go uh as soon as they're announced and you can update to the latest version.

A

B

And I I do have another question which is more about um it. It looks like there is great freedom in how you decide your topologies and how you decide how to set it up. um Are there and maybe maybe they're not documented yet, but are there anti-patterns that we know about and that we should warn people.

C

um Anti-Patents, yes, I'm sure there is a bunch of anti-patterns out there. um The one that comes to mind is just kind of duplicating your yaml.

C

um So if you've got like multiple branches that you're using for like Dev stage prod, um we recommend using something like customize to have like a base description and then overlay with the specific differences between each one. If you find yourself duplicating yaml in multiple places and having to go to multiple places to change stuff, I would consider that Mandy pattern.

C

A

It's great you mentioned earlier, because I also wanted to link a very nice book from our friend Christian Hermann, who wrote that wrote. The patukey tops is a book where he mentioned all those best practices like avoiding multiple yamos, and uh if you are interested, you can download this book from for freeform developer.com.

C

It's a very good book. You should definitely read it. It's only about 45 pages, so you can get through it. Nice and speedily. My copy I have on my iPad is covered in notes. It's it's brilliant yeah.

A

Yeah I really recommend it to all to download it and talking about books. If you want, we have another book, called The get and get UPS another book talking about get UPS on openshift. Specifically, uh you can download it for free uh always from there uh from from the portal, and we have another book called The detox, cookbook.

A

But also by me and my colleague Alex we're sending the last chapter, so it should be really available soon and we use, for instance, only the external secret example. Vault example: algor allows or product delivery so that uh that was a part of the receipt uh there. There will be published first, one.

C

And you can get Early Access chapters available already Tanya exactly.

A

You can go to the website, get there the access chapter and read it before we, we publish it it's going to be available for free on the website, but in the world you can download the the raw edits with all the a syntax error. Of course,.

A

Well, that was great all right. Thank you very much for this session. uh We we learned a lot to think about cleanups. A lot uh also Andrea was not very familiar.

B

With the topic I think now you are super excited and that was very informative to me and obviously, I will go through most of those links. I started reading a couple of books, but uh um I still have to put it in practice. So.

C

uh Pasta get UPS is a great introduction. It's a good.

A

Place, yeah yeah a very good rate. You have the books on there, you have the.

B

A

uh Self-Paced Labs on our developer that I come link I put in the chat before. uh Let me put again uh so if you want to learn about kidoks, we have some some love some tutorial um and if you want to try an open shift, as I already mentioned, you go to the operator Hub and uh and then you can uh try yourself uh by installing it, basically uh that that was great. Thank you very much for joining today. It was a really a pleasure to have you and I hope.

A

You tell you to have you for another episode for another updates on the on the pitops that'd.

C

A

Thank you very much Andrea. uh We have some appointment before we close uh and also announcing we're moving to bi-weekly, so we're not being here next Wednesday we're going to be the the next Wednesday uh and uh if today we have the usual show on openshift TV this afternoon for for us here in the near uh the level up hour as an admin and- uh and we come back in two weeks always Wednesday morning- 9 A.M a bridge time or 10 a.m.

A

uh Central uh summer time uh it was really a pleasure to have you all here. uh Thank you for everything and talk soon. Ciao.

C