Ep#162 What’s NEW in AWS Cloud Intelligence Dashboards (CUDOS)?

Host: Jon

Hi everybody and welcome to the Jon Myer podcast. This podcast is all about our guests from around the world and highlighting their passions in tech and the community. Our topic today is what are the cloud-intelligent dashboards and what's new with them? Our guest today is Yuriy, who is a principal technical account manager, at AWS. He helps his customers build highly reliable cost-effective systems, and achieve operational excellence while running workloads in AWS. Please join me in welcoming Yuriy to the show. Yuriy, thanks for joining me.

Guest: Yuriy

Yeah, hi everyone, and thank you Jon for inviting me. Happy to be here. So let's have a chat about FinOpsand all around and yeah, I'm originally from Ukraine, but at the moment based out of Luxembourg.

Host: Jon

Well, Yuriy, before we jump into the CUDOS dashboards, how about we give the audience a little bit of backstory about yourself?

Guest: Yuriy

Oh yeah, absolutely. So I am, at the moment I'm with AWS and have been here for a bit more than four years in the role of principal technical account manager and my, let's say, day-to-day job is to help enterprise support customers optimize everything on AWS and to make sure that they run workloads in an optimized way according to our best practices and well-architected framework. And cost is a very important part of this, right? And the cost is a concern to every customer. Cost also reflects many other architectural parts of workloads and it happened to be that cost became my passion and optimizing the cost became one of my kind of pet projects, which we together with a couple of other passionate Amazonians launched around three years ago, which turned out to be what we know now as cloud intelligence dashboards or CUDOS dashboard and keeps growing and keeps being used by many, many, many customers. Before joining AWSI worked in different other companies. Mostly was helping some data-intensive distributed systems in the role of systems architect or software developer and before that also ran DevOps teams for mission critical workloads to telecom operators.

Host: Jon

Well Yuriy, let's jump into the cloud Intelligent dashboards. What are they?

Guest: Yuriy

Well, cloud intelligence dashboards is a set of open source templates in Amazon QuickSight, which every customer can deploy in their accounts and get ready to be used actionable, insightful dashboards on top of their spend and usage, which helps them to, first of all, to get great visibility on top of their cost usage and different other operational aspects of their workloads like performance, full tolerance and also cloud intelligence dashboards allow customers to customize the dashboards because we use Amazon QuickSight and everything in QuickSight, you can share within entire organization and you can customize. So those dashboards allow you to build on top and fine-tune the solution to your unique business need, maybe to build some CEO dashboards or to build some different dashboards for different personas within finos because we know that one size doesn't fit all. So that's one of the main interesting things cloud intelligence dashboards have, Is that the name of the project? Yes, so it's how we call it. Within the project, we have several dashboards like CUDOS, our most popular dashboards cost intelligence dashboards and KPI dashboard trusted advisor, which is also known as Style Compute Optimizer. So we have many, many different dashboards, purpose-built dashboards which actually help customers to visualize, get enhanced visibility, visualize their workloads, and optimize them as well.

Host: Jon

Alright, Yuriy, I was under the impression that the cloud intelligent dashboards, which was also referenced as CUDOS, but it turns out that CUDOS is just a dashboard underneath the cloud intelligent dashboards. Is that correct?

Guest: Yuriy

Yeah, absolutely. That's correct. And we started with cost intelligence dashboards more than three years ago. Then we, on top of that, and one more technical account manager from Luxembourg More, we built together CUDOS, where we added a lot of operational insights made for customers who'd like to dive deeper into their cost per resource, cost per cost per some API operations. We went very dive deep, very, very deep into details and we named it CUDOS and it became very, very popular. So we have a lot of customers who use it. We have a lot of technical account managers who use it internally with customers as well. So that's why it's common. People know us under the name CUDOS because it's the most popular dashboard, but we have more than this approach of creating and pre-building templates for the customers which customers then can deploy in their accounts.

Guest: Yuriy

This approach became very popular and then we extended it. We've built, for example, dashboards that visualize as trusted advisor data, which is called, and we also have a dashboard for Compute Optimizer and keep expanding this approach to even further domains. So we have now a dashboard for sustainability, and we have a dashboard for cost and usage from Azure on AWS. So we see that customers are allowed to have a single pane of glass for their cost usage and everything cloud-related in Amazon QuickSight and use QuickSight as their go-to tool to share these reports and dashboards within the organization.

Host: Jon

The cloud intelligent dashboards were created three years ago or the name and everything was created three years ago. What was your first dashboard created? Was it the CUDOS or did you have a different variation of it and let's talk about how it's progressed?

Guest: Yuriy

Yeah, we started with the cost intelligence dashboard and it was created by Ali Whitman back then. Then Cost Intelligence dashboard gives customers great visibility on higher levels. So it's usually we see that more like financial or executive people use it. Then we took a cost diligence dashboard together with my colleague Timor and built a CUDOS dashboard on top of it where we actually as a times we added a lot of insights which we used to present customers during our management business reviews, and cost optimization reviews. So actually we've put pretty much all the knowledge and experience that Toms have in the field into the dashboard. So every customer can have this as a self-service tool. So we came up back then with the name of CUDOS, we tried to play with cost and usage dashboard and then there was an idea around then we want to hear CUDOS for CUDOS and we came up with this

Host: Jon

Name, I keep using the pun, CUDOS to you guys just because of the name. I think it's cool. I think it fits everything that you guys are trying to do. One of the questions I have is you created these dashboards three years ago and you're working because back then when everybody was going to the cloud, cost wasn't a factor. We weren't thinking about it. But in recent years, FinOpshas come around, it's been very popular, the culture, everything's been shifting towards that cost in your cloud has goinggone up and understanding it that it does cost rather than having a traditional data center, but if you use the cloud for what it was meant to be, then your costs won't be as much as you originally thought. My question for you for these dashboards is now that they're three years old, is there ever a thought to have them as a standard by AWS that's automatically deployed within your account for visibility?

Guest: Yuriy

Well, so we still keep innovating, yes. So we developed them three years ago, but during this time we were adding new visuals, we were adding new capabilities to existing dashboards, we were adding new dashboards and we based at the moment as in AWE as well architected lab and we are part of AWS well Architected Labs. Customers can easily deploy these dashboards by running cloud formation templates. Deployment takes usually around from 20 to 25 minutes. It's very easy and simple and we work closely with our service teams who run the billing console and all experience around and work very closely on the next steps or how to bring these dashboards closer to the customer. So this work is in progress, I would say, but overall it's for customers, it's really about the deployment part of one-time operations, they deploy it once, but then they have it all in QuickSight and then they keep using dashboards in the QuickSight and they share within their engineering teams, product owners and other FinOps related personas. So it's the deployment part I would say it's not the most critical here. It's like the next usage of the dashboard is more critical, but we are part of Architected Labs now, so let's say It's pretty much official.

Host: Jon

Well, I like that, Yuriy. So let's talk about that. You're a part of the Well-Architected Labs. This is a newly released landing page that you were posting on social media. What's it been three weeks now? Do you have some stuff you can share with us?

Guest: Yuriy

Oh yeah, absolutely. So as I mentioned, we are part of Well Architected Labs and let me present here a bit to share screen in Well Architected Labs. We have labs for different pillars of well-architected framework as you can see here, operational excellence, security, reliability. We used to be in the cost optimization section or a pillar of well-architected Labs, but with time we have a bit overgrown a cost section because we have, for example, a trusted Advisor dashboard, which shows also recommendations and risks for security, for full tolerance for performance. Now, we are working on the very least sustainability dashboard, which covers the sustainability domain or sustainability pillar. So together with the well-architected team, we decided, hey, we pretty much can be hosted in a top-level section because we cover visibility for other well-architected pillars. We try to uncover how to identify what we are talking about in well-architected frameworks or white papers, how to visualize it all, and make sense of it.

Guest: Yuriy

So that's why we now moved to the top-level section in Architected Labs. Here you can find US cloud intelligence dashboards and this page acts as a reference page for all our dashboards. So as a reference, what I mean is the actual deployment guides could be hosted on other different platforms. So you can see here's a lot of links. So you can have a look at our demo dashboards. So, here's an example of CUDOS also you can go to the details page for example and read more about the dashboard itself, like some details, descriptions and go to the deployment guide, and then follow the deployment steps, on how to deploy the dashboard. And as I mentioned, it's pretty easy. There are a couple of cloud formation templates that you need to run and get it all deployed. And for some dashboards here we have references to other workshops like sustainability proxy metrics, it's hosted in sustainability, well-architected labs. So if you go to deploy here, you'll be actually in well architected sustainability pillar and the dashboard is located there. That's why we decided that it makes sense here to have this reference page with all the links. So it's like a dashboard portfolio for customers, in which they can choose which dashboard to deploy, and which dashboards to have a look at in a demo and make a decision later. We also tried to target links to a target audience who might be interested in those dashboards.

Host: Jon

I was going to ask you why the cloud intelligent dashboards were part of the well-architected labs and I can understand why based on the deployments that are happening, the dashboards that you have available, the sustainability, and the cost optimization, you going into operational excellence. I can see the dashboards as part of the well-architected labs, but will they be part of the well-architected framework at any point or they're just there available for people to utilize test out, and get a feel for what they're going to be like in their environment?

Guest: Yuriy

We are already part of the well-architected framework, so cost optimization, well-architected white paper, and recommendations that you have in the tool reference cloud intelligence dashboards in some sections. So we are already part of well well-architected framework. So for example, when we talk about cost visibility there are different sets of recommendations. So there are recommendations about the deployment of cloud intelligence dashboards to get this very detailed cost visibility.

Host: Jon

I like it that the well-architected framework is referencing the dashboards that it's a suggestion, it's a recommendation to deploy these out within your environment to get full visibility into the cost of any type of workload based on the well-architected review process.

Guest: Yuriy

Absolutely, yes, that's a goal. And also in the dashboards we try to visualize and help customers to simplify the reasoning about what we talk about in well-architected white papers and frameworks as well. Sustainability here is also a very interesting example because well architect in sustainability, well architect at whitepaper, we talk about proxy metrics here, how to assess your sustainability improvements and we describe those pro metrics, but in the sustainability dashboard, we visualize those prometric so customers can now connect the white paper to their data.

Host: Jon

Yuriy, talk to me about the latest things that are happening in the CUDOS dashboards. Will they ever be a part of TAO?

Guest: Yuriy

Yeah, so let's have a look. So there are a couple of recent interesting releases that we've done in I would say previous couple of months. So let me switch to the demo dashboard and let's go to Tao. It's our dashboard for trusted advisors. It's available for customers who own business enterprises and OnRamp support because simply the trusted advisor is available for those customers. And here in the dashboard, you can find a summary tab for your high-level overview of trusted advisors' checks across different well-architected pillars like security, full tolerance, and performance cost optimization with the number of flagged resources with monthly, monthly progress towards mitigation. Also here we have separate tabs like details about security and resource level granularity for different checks. And the same for us for cost optimization. You can see what are your idle database instances and what the estimated savings, et cetera.

Guest: Yuriy

But what we recently included in the summary tab is customers are often asked to have the ability to dive deeper into trends in a single place without jumping between tabs. So what we added here is, the so-called Trusted Advisor Explorer feature we call it. So here on the right side, you have all your linked accounts with monthly trends of flagged resources. On the left side, you have all the trusted checks with the amount of flagged resources and you can easily understand, okay, look, here is some account that had like 574 checks in February. You can interactively click on it and then the rest of the visuals will be filtered based on your selection. So now you looking into that particular account and as you can see there are some different security recommendations like cost, et cetera, but you can find, hey, here's some recommendations or checks which have the most resources here. So you can click on the check as well. And then here detailed view will be also filtered based on your selection and you will have all resource IDs, all accounts, and all regions. So you can easily then know, okay, here's what I focus I should focus on. Maybe you can go later to the security tab and view more details or view more trends. But this is a single place that allows you to dive deeper into pretty much any trusted advisory check and resources with resource ID which is concerned, which is flagged by a trusted advisor.

Host: Jon

I can see this as very powerful for everybody for visualization within a single place. I think the trusted advisor explorer caters to the FinOps community and those who are trying to visualize everything and just be able to accomplish those tasks or to see where they can kind of tackle some of those challenges within their cost optimization within their environment. This is awesome that you guys put this together

Guest: Yuriy

As I mentioned, we keep adding features to our dashboards and customers can update dashboards if they have deployed previous versions, they can run an update and get the latest version with all those new features.

Host: Jon

Well actually talk to me about the update. How does that process work? So I've deployed this dashboard following the well-architected labs and the deployment guide, but now you guys put out a new update, and I want those features. Is it a rip down or is it just an add-on that updates the existing deployment?

Guest: Yuriy

Yeah, it's as simple as running three commands in your cloud show. So it's like you need to install our Python module C-I-D-C-M-D and then run the C-I-D-C-M-D update. It'll discover which dashboards you have at the moment. We'll discover which dashboards pending updates where we have the latest version of the template which is not available, which is more recent than the version of the dashboard that you have deployed you just pick the dashboard from the list and you get it updated. Or you can watch the video here where we cover this step by step, but it's super easy and customers always have a way to get the latest version of the dashboards and get all the new features we should provide.

Host: Jon

Yuriy, talk to me about FinOps a little bit and that different personas want to see different variations of it from the engineering to the finance to the CIOs to the business managers. You were talking about how QuickSight and the cloud intelligent dashboards cater to those different personas that are available to it. Is it as simple as a dropdown and I want to select certain personas that I've already set up, or is there an additional configuration that needs to be done

Guest: Yuriy

Once you deploy the dashboard? Yeah, you have it pretty much in this state as is. Yeah, you can share it in different ways, which we different mechanisms that QuickSight supports. For example, you can share it via the identity center if you use things like SSO or you can even share the dashboard via email. So in many cases, we see that some FinOps personas usually don't log into the AWS console, especially finance people or executives or some finance personas like product owners or maybe engineering leads. They need visibility over just a group of accounts. And if you look at what you can do in the AWS console, you can either see everything if you log into the payer account or you can see the account per account basis and you need to log into every account. So here we consolidate all the information. By the way, in the dashboards, you can consolidate data from multiple payers in one single place.

Guest: Yuriy

So it's like you connect your all AWS organizations in one single place and then yeah, you can share a dashboard, you can apply QuickSight role level security where you allow your engineers or allow your other FinOps personas to see only data which they should see. And you map linked accounts to your QuickSight users. It's very powerful because it enables these use cases where you want to share some group of accounts with someone, but you don't want to share entire organizational data. And then what else you can do is in QuickSight you can create, a so-called analysis or editable version of the dashboard. And there it's just you can do it via just a few clicks like you click save s and you create an analysis, and there you can see all the columns, all the fields which we use, all the logic which is used here.

Guest: Yuriy

You can reshuffle and change the position of the visuals, you can create a new tab, for example. So it's pretty much easy to do because it's in QuickSight, which is a BI tool, which is built for ease of creation of this kind of visualization. So many customers build on top and they learn how we calculate particular things like savings in discounts in a single place for all your kind of savings plans, sports, all refunds, et cetera, in one single place. So customers learn how to calculate at all because it's pretty much not very trivial. Yeah, it's quite challenging. So that's why they love and that's very, very, very common pattern is to take these dashboards and customize and build on top.

Host: Jon

Yuriy, I can see the power at anybody's fingertips to customize the dashboards for their specific needs and visualization. Is there ever a possibility that there's just too much data for them to realize? I mean, I can see the value in all the data, but at one point, I don't want to say analysis by paralysis, but there's just too much data that it's hard to understand what you need to do next, what you need to tackle, what's driving my cost up.

Guest: Yuriy

Yeah, absolutely. And that's why you can call sometimes CUDOS a supermarket of the visual. So visual visualization so you can find the ones that matter for you and just move them into separate tops, which makes sense for you, which matters for you. So that's what you can do. Another thing is, in this case, I would recommend also exploring our KPI dashboard because KPI dashboard allows you to, provide you set of predefined cost optimization KPIs like the percentage of your workloads running on previous generations of EC2 instances or the percentage of your workloads running on SPOT or on graviton or amount of volumes EBS volumes which use GP3 are the most cost-efficient EBS volume types and the interesting part that you can set your own goals here. So if you want to migrate RDS to graviton, you can set your goal like 90 or 100%, you might want to set EBS to GP3migration to 90%.

Guest: Yuriy

You can adjust your goals here and then you can see pretty much in a very structured way, you can see where you are, what's your monthly progress, what's your goal, and whether you achieved it or not. So it allows you to simplify this reasoning about, so it's pretty much straight to the point what you should track and where you are with this. But another added value is that we also estimate potential savings, which you can get if you achieve the goal. So for example, for different goals, you can see what you can save or what you could save in previous months. So first of all, you can easily understand, okay, so we optimize it by around seven K, our potential savings, so it's good. Another thing is we can prioritize the efforts so we can see what is the most impactful change and also assess how hard it would be for us to implement. So for example, migrating to sport might be not the easiest exercise, but migrating to GP three might be much easier. We can see that migration to GP three brings a pretty decent amount of savings for this demo customer. So that allows you to prioritize the efforts as well.

Host: Jon

Yuriy, this is very clean and powerful. The KPI dashboards I think is something and I recommend anybody deploying it up because it's one of those hidden costs or one of those hidden things that you're not visualizing or can't visualize. It's very clean for that. Let me switch gears a little bit. What's next for CUDOS? What's happening?

Guest: Yuriy

Well, yeah, so next we keep adding new functionality. So we are exploring a lot of additional visuals that we want to add related to analytic services, including QuickSight for example. Another part is we are going to also work on some performance optimization of some visuals to optimize them to be loaded very fast regardless of the size of your cost and usage report. As I mentioned, customers often combine multiple payers in one single place, and sometimes for very large customers it might be loading a bit slower, so we are going to optimize performance. So regardless of your size, it'll be super, super fast and the performance will be stable. One more thing, what we are going to release also is, that we are working on a couple of new dashboards. So we will release, I think shortly a new dashboard. So stay tuned. Also, we are working with service teams to make this experience part of the standard service offering. So that's kind of our part of what we are doing. One more thing is that we recently joined FinOps Foundation as AWS, and we also joined the Focus group for open cost and usage specification. So I think there's a lot of potential there as well to align to focus terminology, what we have in our existing dashboards. They also contribute to focus specification and then use specification in the evolution of our dashboards as well.

Host: Jon

We already, speaking of joining the FinOps Foundation, reinvent is coming up with what's happening for the cloud intelligent dashboards and FinOpsat reinvent.

Guest: Yuriy

So we have plenty of sessions and actually, I think we have a whole blog post about that. So let me just bring it to the screen where we have to outline all the FinOps-related sessions or the majority of ops-related sessions. So here you can find it on the Internet and we have quite a lot. Yeah, we have breakout sessions, we have workshops, and they mostly around the entire AWS service offering for FinOps and cloud intelligence dashboards specifically this year we have a workshop, COP 312 track and visualize the cost-effectiveness of your workloads and KPI targets where we will be providing participants a hands-on experience. So they would be able to use the dashboards and customize the dashboards in their accounts and build business unit metrics by combining KWS cost and usage with some business metrics like cost per customer, cost per transaction, and things like that. And bringing this all together. So in QuickSight, it's quite easy to do. So we will be running this kind of lab during this workshop and also we'll have a breakout session in architecture track COP 319 where we are going to talk with Mike Draft from Dolby about the impact of customer culture and FinOps in general on the overall health of customer workloads and architecture and other well-architected domains. And we also might have some guest speakers there as well together with us. But let's see, closer to the dates,

Host: Jon

Yuriy, how many sessions or workshops are you doing at this year's Reinvent?

Guest: Yuriy

So yeah, I'm participating in the workshop and also the architecture breakout session, but also involved in helping some other folks. I think what I like in preparations to reinvent is that a lot of FinOps-related professionals and enthusiasts come together and help each other on different topics and different sessions. So if someone has some particular part of KPIs or something like that. So we work together with everyone who is involved in that and help each other to make content better, to just bring more diverse experience or diverse opinions on the feedback. And that's what I really, really like in our preparations to reinvent that. We act as one big team of all the field FinOps specialists who then come up with the best possible content for our customers.

Host: Jon

Yuriy, are you looking forward to one specific session that either you're presenting or somebody's presenting at Reinvent that you'd like to highlight for the audience?

Guest: Yuriy

Well, look, I would suggest picking your session as I will be. I cannot attend sessions, so I'm not considering them from a standpoint of what could be useful or not. I would say for every customer, it's better to assess where they are in their FinOps journey if they're looking for more kind of hands-on or more practical advice, which they can take out of the session and go and implement. Better for them to attend workshops or check talks or something which or builder session, right breakout sessions would fit for maybe decision makers for visionaries, for more kind of people who are involved, not in hands-on building, but rather deciding how you're building stuff and where you're leading to. So for those people, I would recommend joining breakout sessions, and as I mentioned, we have also session A or C 319 in the architecture track where I will be presenting as well. So I would say choose your right session because sometimes it's hard to navigate, hard to pick out of all the sessions which ones you want to attend. So I would say, yeah, just decide based on what experience you want to get.

Host: Jon

Yuriy, I'm going to do another, a little plugin here about our faces in FinOps podcast that's happening at the Blue Wire Studios, Yuriy and the entire team, we're going to have them in there talking more in-depth about the cloud intelligence dashboards, CUDOS, what's coming up, some of the releases, some of the events that are happening. Stay tuned for that. Yuriy, are you looking forward to this awesome adventure out in the Blue Wire studios?

Guest: Yuriy

Oh yeah, absolutely. And I think it's going to be great because we have such a cool lineup there.

Host: Jon

Yes, don't miss this. A recording is happening on Monday. We're going to try to get it out as quickly as possible. It will be released sometime in December, but stay tuned for that episode. Yuriy, I have to wrap things up, but thank you so much for the in-depth conversation around not only the dashboards, CUDOS T, the KPIs, and everything that you're releasing, but also what's happening at Reinvent because I'm looking forward to it.

Guest: Yuriy

Yeah, thanks. Also, thank you for this opportunity to speak about cloud intelligence dashboards, and yeah, looking forward to meeting everyone at Reinvent, and joining our sessions related to FinOps also when you pass by our CFM or FinOpskiosk, just pass by to say hi. Most likely when I'm not on the sessions, I will be somewhere around there.

Host: Jon

Yeah, definitely. Thank you, Yuriy, so much for joining me. I appreciate it.

Guest: Yuriy

Thank you.

Host: Jon

All right, everybody, this has been the Jon Myer podcast. Don't forget to hit that, like subscribe and notify because guess what, we're out of here.