Transcript 0:00 But if you tell that, "Oh, we managed to reduce cost per customer by, you know, by, by, by 10%," oh, well, that's kind of, that's, that's completely different story for executive. 0:10 One of the key ideas that we have is that we... 0:14 Like, what, what we want to actually, um, achieve with these dashboards is to give customers a very, very quick start into connecting them into insights which they can get from the data. 0:25 Uh, they took our dashboards, but they completely built, like, absolutely mind-blowing FinOps, uh, home-baked tool, uh, on top of this. You build it, you run it, but you're also responsible for, um, cost effec- Yeah... 0:37 effectiveness, yeah, or for the cost of this. So it adds more responsibilities and accountabilities, but at the same time, without this, it would be, it would be hard to drive the change. 0:47 So we started to build and, uh, experiment with the visualizations and the dashboards in QuickSight, and we realized that it's, um, actually possible not only build insightful s-stuff, but also to use QuickSight to share the templates with, um, every single customer. 1:04 For, for FinOps community also GenAI brings quite interesting challenging can, um, challenges not just, uh, how to use it for FinOps, but also how to track the, you know, the cost of GenAI services, how to optimize them. 1:18 So welcome everyone to a new episode of the FinOps Weekly Podcast. Uh, today we are honored to have one of the, uh, masters of visualization in, in FinOps. We have Yury from AWS. How are you, Yury? Uh, hi. 1:33 Hi Victor, hi Da-Damian. Uh, I'm great. Thank you for having me today. Um, yeah, how are you? Doing well, doing well. Uh, yeah, an honor to, to have you here. And today we also have our co-host, uh, Damian. 1:47 Damian, how are you? Fine, thank you. It's very special, uh, event. I, I, you know that I love the dashboards [laughs] and all the dashboards. Me. So it's [laughs]... I'm very excited about having Yury here. 2:00 Um, so I guess, uh, you know, it's a new year, 2025. I guess, uh, Yury, you can tell us a little bit what, what is going on with the dashboards. 2:12 Are we gonna see anything that you can share with us about what i-it's gonna happen this year? Yeah, absolutely. So we can, we can definitely start from that. And first of all, I would like to 2:25 make a quick summary of what we actually delivered in 2024 in, you know, cloud intelligence dashboards. Um, and it's qui- it was, like, quite a intense year. 2:38 We had quite a lot of, uh, we released quite a lot of updates to existing dashboards such as Kudos. 2:45 Uh, we released the whole five new, um, significant version improvements, added a lot of, uh, new insights and visibility for, you know, FinOps community, such as tracking of extended support, um, more insights for, um, 3:03 AI/ML workloads such as we added, you know, comprehensive visibility for Bedrock, for Amazon Q. You know, the, quite a hot services at the moment. Um, and also some 3:16 significant improvements for DynamoDB visualization for management and governance services. 3:22 And we also released 10 new dashboards, um, which cover different areas from, uh, containers cost allocations to something more general like focus, and also dashboards which go beyond FinOps, health events, support cases. 3:41 So, um, because the concept of getting data and applying BI to a data, it's pretty much, you know, generic. 3:49 It's, it can be used for any domain, so we really expanding now to, uh, more domains, um, uh, such we're going into, uh, operations. 3:58 We also plan to go more deeply in, uh, security to, uh, cover things like Security Hub, for example. Um, yeah, and we also published, uh, four customization guides because one of the interesting 4:16 th-things about our solution is that it's fully customizable and it's fully open source. 4:21 So every single part, it's available for everyone to review how it's built from deployment automation to, um, the calculations on the dashboards, how we calculate anything, how we, uh, come up with all this visualization. 4:34 So we also publishing customization guides because customers can see how it's built, but they can also build on top. 4:42 So that's, um, that's very important 'cause we use these dashboards not only as a solution for customers, but also as a, like, uh, a place where they can learn how they can build similar stuff. 4:55 So we want to, um, like accelerate their FinOps journey as they're, um, uh, start into the data. But, uh, something [laughs] sounds like we had a very intense year. 5:11 Actually it was every time that, that I saw the post on LinkedIn was, wow, more dashboard and more dashboard. And, and- Yeah, I, I, I posted an article, um, before end of year with, uh, our, uh, year-end review. 5:25 So, um, if anyone is interested, you can read. But I hope, um, 2025 will be even more, uh, intense and more, uh, exciting and interesting. We are working on quite significant initiatives as well. 5:39 Uh, so roadmap is pretty full. But, uh, if anyone has some ideas, feedback of what they would like to see new in cloud intelligence dashboards in 2025, uh, drop me a line. 5:51 Would be super happy to hear that, to discuss with anyone. Excellent. Okay. Great invitation. So everybody is open to share with you [laughs]. So come on, guys. 6:03 And, um, but this also, uh, you know-I always wanted to know how all this started and how you, how you got into this. You're now leading all this, uh, dashboards initiative, which I love, as I mentioned before. 6:19 So it's very interesting to-- for us to know how, how it all began. Yeah. Uh, that's, that's a great question. Um, not often, uh, I'm asked about that, so super happy to talk about this. 6:30 And I want to start, you know, with recognition of the, of the team, because all these deliveries that we, uh, that we've done in twenty twenty-four and b-before it, all the hard work of the team, 6:42 uh, and all the contributors, all the authors of the dashboards and, you know, who actually very, very cus-customer obsessed Amazonians who want to bring basically the knowledge, um, which we have, uh, as SMEs to every single customer as a self-service tool. 7:01 It all started with, um, I think in twenty twenty, uh, in around... 7:08 At, at the end of the twenty twenty, um, it started as basically several ideas, um, between a few technical account manager-- uh, managers like me and, um, 7:25 back then my colleague Timur, and we also partnered with Ellie Whitman back then. 7:30 So we started to build and, uh, experiment with the visualizations and the dashboards in QuickSight, and we realized that it's, um, actually possible not only build insightful s-stuff, but also to use QuickSight to share the templates with, um, every single customer. 7:48 And when we started to show our visualizations to, to our colleagues, to customers, the first question was, "Oh, it looks great, but how can I, um, how can I get it? How can I use it, uh, with my customers?" 8:02 So that's where we realized that we need to shape it in some, you know, form of product. 8:09 And we released two first dashboards, uh, Cost Intelligent-- Intelligence dashboard, which was focused more on like high-level insights, 8:18 and then following with Kudos dashboard, which was, uh, taking this high-level insights and also enriching with a lot of operational details. 8:28 So that's where we went into, you know, a lot of resource level visibility, what-- which resources you can optimize, not just seeing the trends, but also like getting into resource IDs and all the details. 8:40 And this is very, very well resonated with, uh, internal communities because, um, our dashboard is also used internally, and we've got overwhelming amount of feedback, which actually is reuse until, you know, until today. 8:59 So one of the key parts of our solution is that we take a lot of field feedback and we iterate, we release new versions, we, uh, we try to elaborate all the new ideas which appear. 9:12 So basically, we started to incrementally add new features, new ideas. Uh, then we started to receive feedback from service teams, you know, something like, uh, 9:24 visuals which show you inventory of your ODCRs or, uh, some visuals which appear like for backup and, uh, storage. 9:34 They come from, uh, different service teams, uh, from solution architects who work deeply with service teams. So it's 9:41 so many knowledge there and which we actually were able to bring to any customer as a self-service tool, uh, so they can deploy in, in, in their accounts. Uh, yeah, and since then, we had 9:58 a lot of work with many, many customers also getting their feedback. 10:05 And by the way, uh, I just want to thank to every customer who, you know, sharing the feedback about how they use our dashboards, how they, um-- what we could improve, what is missing, because we are always open to hear this feedback. 10:18 It's literally gold for us. Um, and yeah, so it's like after, I think, uh, four and a half years of, uh, you know, uh, like i-incremental improvements, uh, of this, we are where we are. Uh, and 10:37 one of the key ideas that we have is that we, uh, like what, what we want to actually, um, achieve with these dashboards is to give customers a very, very quick start into connecting them into insights which they can get from the data. 10:54 So because every customer can enable cost and usage report, every customer can, um, get this data. But digging into raw data is very, very challenging, believe me [chuckles]. 11:04 Like, um, uh, I-- as a person who like build a lot of this, so, uh, it's, it's quite hard. 11:11 So what we want is to just simplify a start for customers to get in, you know, to get them this, um, tooling where, uh, they don't need to pay, you know, any license fees or things like that. 11:22 Just quickly get up to speed, uh, deploy these templates and, you know, pay only for underlying services which they used, um, and, and then see what they can build on top. 11:34 Because we have a lot of examples like from BMW, uh, where we published a joint, uh, case study where, uh, they took our dashboards, but they completely built like absolutely mind-blowing FinOps, uh, home-baked tool, uh, on top of this. 11:48 So I'm super expired from th-- inspired from, uh, this kind of, um, stories as well. So, uh, yeah. 11:56 And then just want to say thanks again to our entire team, you know, who, who helps us to drive this, uh, forward, uh, over these years.Um, uh, so, uh, yeah. 12:10 And as you can see, now ideas exploding with the new dashboards like [chuckles] the la- the last one which we released is for Amazon Connect, a very specific dashboard for, um, customers who run these call centers and use Connect. 12:23 And, like, it's apparently to... For a single service, you can come up with, like, four or five tabs of, you know, visuals with- Wow.... 12:33 uh, with insightful visualizations which, which can help customers to understand what actually we're paying for, you know, what we can do with this all. Th- these are very good points there because I also... 12:46 I, I use tools and, and the, the cost, uh, cloud intelligent dashboards also for... Like you say, you can 12:53 ramp, ramp it up, install it very easily, ramp it up, have an environment set, and then, uh, you can copy it as an analysis and do changes because it provide you, 13:04 uh, all the information very well architectured and very well, which... And I think it's very, uh, excellent. And for those who are listening also, they can look at all the dash- dashboards online in the workshop, right? 13:17 Maybe we can share afterwards the link. And they can play and see what the information, uh, and all the, all the dashboards and all the information they have. Yeah. Uh, which is very, very, very good. And, um, wow. 13:32 [chuckles] Very excited to know all the, this information. Right there, Victor? Yeah. 13:37 Yeah, and it's, uh, what, what I was, uh, thinking while, while you were saying about the, yeah, all the, all the news and all the, all the, the goals of the, of the dashboards is, like, we are seeing, uh, a lot of all the trend of, you know, moving the, the insights from the, like, cloud data, uh, information to, uh, the business value. 13:58 Mm-hmm. And here the, the main thing is, you know, translating that information to KPIs and to OKRs and all this stuff. 14:05 So how do you think this, um, you know, uh, even the dashboards or in general the, the information that you can provide on this, uh, visualization tool should connect to, to the KPIs? And which... 14:17 If you have some, like, what are the KPIs or what are you seeing from the customers that which KPIs are they, they're providing them the most value? Mm. 14:26 I, I assume it's mostly depending on the case because each business has its own KPIs, but I assume there are some, some generalities there. Yeah. Yeah, absolutely. 14:34 It's also v- v- very good question because it's, you know, FinOps is not only optimization, you know? Yeah. 14:39 It's also being able to understand what, uh, business value you get out of the, out, out of your investment to the cloud. So, um, I usually when I think about KPIs, um, I usually split them into three, uh, categories. 14:55 Mm-hmm. So, uh, we can go from, from bottom to top. Uh, first is like, uh, cost efficiency KPIs. 15:03 It's basically KPIs which build around how well you are leveraging one or another cost, cost efficiency or cost effectiveness opportunity. 15:13 For example, um, uh, you know, like, uh, how, uh, what portion of what percentage of our EBS volumes is running on GP2 out of all general purpose volumes- Mm.... or out of all volumes? 15:27 Or what is our savings plans coverage? Or, um, what is our percent of EBS snapshots which are older than one year? Or, uh, what is our percentage of, uh, S3 storage, uh, 15:42 which is standard storage over, you know, entire, um- Yeah... S3 storage? So basically very, very 15:49 low level, very technical, uh, KPIs based on pretty much every optimization opportunity or, you know, percentage of right-sizing savings opportunities out of your total spend again. 16:01 So it's like, um, basically those KPIs are quite powerful for, uh, technical teams because if they, you know, look at them, they basically tell what you can do to improve, you know? 16:14 If it's like GP- G- uh, GP3 adoption, you kind of... They guide you. "Okay, you need to, to, to use more GP3," you know? It's like, um... 16:24 Or I don't know, if it's adoption of Graviton, they kind of, again, self-explanatory. "You need to adopt Graviton." 16:30 So it's a power of those cost effectiveness KP- KPIs because, uh, when organization starts to promote them- Mm-hmm... it's, it becomes, it, it, it builds habits within the technical teams. So, you know, if we have, like, 16:46 10, 20, 30 KPIs like that and we follow them, we can, um... It basically signal for engineers what, what we are tracking for. And, uh, here it, what, what is important for, uh, companies to, 17:03 to note that every time when you get some success story, someone improves something, optimize something, every, like, every success story could be turned into KPI. 17:13 You know, if we figured out some new optimization opportunities like, for example, oh, we, I don't know, we have multiple cloud trail- trails which duplicated and things like that, so why do we need it? We optimize them. 17:26 It could be defined as your KPI. Like, you should not have duplicating cloud trail- trails because, you know- Yeah... every second one and every, every not first one is paid and stuff like that. So all... 17:38 Like, always when you optimize something, thi- think about turning it into cost efficiency KPIs. But obviously it would be a lot of them- Mm-hmm... because, you know, uh, they're very, very low level. 17:49 Uh, when you start thinking about business impact of your optimizations, you need to kind of have a bit less KPIs, but more kind of connecting dots and connecting this to business. 17:59 And that's where I come with a second layer, which I call usually infrastructure unit cost. 18:06 So it's when you start measuring the-Average cost, uh, per, uh, for example, gigabyte of data for storage or average cost of normalized instance hour. Uh, and that's 18:22 a higher level measure which can help you to assess, okay, so what's our cost of infrastructure over time, yeah? And, uh, when we optimize something on this lower lever- level KPIs, we should see reflection into 18:38 infrastructure unit cost, right? And vice versa. If we see that infrastructure unit cost increasing, then we can... 18:44 It, it gives us a signal of how, uh, that we need to dive deeper into this, uh, cost efficiency KPIs and understand why our infrastructure unit cost increased, yeah? Mm-hmm. 18:54 And the beauty of infras-infrastructure unit cost that you can build those, uh, metrics or KPIs, um, just out of cost and usage data. You, you don't need any cert... Kind of your custom metrics or something like that. 19:07 So you can pretty much build it by divide, by dividing, uh, cost by usage, right? So it's, uh, it's very easy. And we have these kind of things in cloud intelligence dashboards. 19:19 We allow you to see in, in infrastructure unit cost in many, many dashboards. And the third layer, it's where you actually can measure the impact on the business. It's, um, business unit cost, right? So it's- Mm.... 19:33 where you go into this so-called unit cost economy, where you start tracking the cost per your business drivers, such as cost per customer, cost per transaction. 19:42 So that's where you actually connect, um, all the infrastructure to the business, and those metrics are usually very well under-understood by, um, by executives, by financial people. 19:54 Um, you know, because if we tell someone we, I don't know, optimized S3 storage by twenty thousand per month, what will it tell to, you know, like business people who think about revenue, who think about margins? 20:08 It's like, like it, it... Was it budgeted? Was it like forecasted? Was it planned? Like, what, what, what, what does it mean to our business, you know? 20:18 Like, uh, but if you tell that, "Oh, we managed to reduce cost per customer by, you know, by, by, by ten percent," oh, wow, that's kind of, that's, that's completely different story for executives. 20:31 Um, yeah, so that's, that's, that's, that's the layers of KPIs and, um... But, but to build the infra-- business unit cost KPIs, you, you, you need to be able to join your business data and, and your, um- Yeah... 20:45 infrastructure data. And that's where it comes very handy to use, um... Because you, you cannot bring your business data to default AWS, um, Cloud Financial Management service. Mm-hmm. 20:59 So that's where it's very handy comes, uh, things like QuickSight, um, where you can, uh, take the existing QuickSight dashboards and join the, like bring your own data there, because like QuickSight supports a lot of data integrations. 21:12 And, um, and actually, and, um, add your business metrics to existing dashboards. Uh, so it's very, very powerful. Actually, at re:Invent, we, we, we had a workshop, uh, this year where we, where we covered this. 21:27 So basically, how to measure impact of your-- business impact of your optimizations and where we like step, step by ste-step everyone built during the workshop this, um, business unit cost dashboard, uh, by joining together cost and business data. 21:44 Oh, that's, that's super cool. I think the classification was great. I mean, uh, I think it's super valuable to have this clarified and separated. 21:53 And also, yeah, like you can do these amazing connections with, with the, the tools that, you know, the providers have, in this case AWS. 22:00 Like you can mix a lot of things, and I know, uh, Damien knows a lot about QuickSight and all the combinations that, that you can actually do there. 22:08 Um, so moving on a bit on like, uh, these deep dives and these connections, I want to know a bit... Like now we, we have addressed the business part, and now I want to move a bit on the engineering part. Mm-hmm. 22:20 So we have this, uh, the dashboards, but these dashboards could or could not be, you know, managed by engineers. Uh, so let's say that these dashboards are managed by a central FinOps team. 22:32 How do you think the, that team or the, the company in general should translate this information to the engineers, and what kind of, you know, cultures or what kind of behaviors, uh, should be implemented to help the engineers actually make the, the optimizations there? 22:51 Yeah. It's, uh, you know, like million-dollar question, basically. [laughs] Yeah. [laughs] That's why I ask it. [laughs] Yes, yes. So, um, every organization approach it in a different way. 23:03 I would say first question should be, why would engineers ever go to this dashboard, you know? Mm-hmm. Engineers have their, you know, like operational metrics dashboards. They have, um, their performance dashboards. 23:19 They have some latency, some maybe availability dashboards, and so many, so many things they usually care about, and then we, you know, we bring into picture cost. 23:31 Like I would say we need to answer a question, why would engineers prioritize this over everything else they have? And that's where cultural, uh, aspect comes into play. 23:41 So it should be very clear for engineers that it's, it's priority. Yeah, so- Mm. Because if organization recognize something else, you know, like business grow, uh, growth- Mm-hmm... um, I don't know, maybe availability. 23:56 If, you know- Yeah... it was a problem of customer sentiment or something like that. Mm-hmm. 24:01 If, if organization embrace or, or recognize those things and cost is never in, you know, highlights and recognition, that will build a pattern where engineers will focus on other things. They, they, like it, it... 24:14 They wouldn't have intent. 24:15 But if they will see that you actually recognize cost awareness, you recognize, you know, this frugal mindset-Um, a frugal approach to architecting, uh, things that, and actually like the frugal architect principles, uh- Mm-hmm... 24:31 of, uh, by Werner Vogels. It's like they, they're just amazing. So if you start as organization more highlight these kind of stories, uh, the things recognize people, 24:42 it first of all, it will give them a hint, oh, like we need to do more of this. So once they have some motivation, you know, to go to this dashboard, then that's place where you actually need to, um, find some champions. 24:58 Yeah. Mm-hmm. So like you need to find the champions within the different engineering teams and have like, um, you know, like a squad of, uh, FinOps ambassadors, FinOps champions, frugality 25:11 gurus, frugality masters, like whatever you call them. Uh, but you need to build a network of 25:18 very technical people who, you know, work and understand all the depths, not just of, you know, AWS pricing structure and stuff like that, but also like how applications build. 25:28 Because sometimes to improve, uh, to, to optimize cost, you, you actually like c- need to optimize your code, you know? So like you need to- Mm-hmm... 25:38 to build this network of this, uh, technical people who can speak with, um, who, who, who understand every breadth and depth of, of, of their applications. 25:47 And, and then leverage these technical people through, you know, actually championing your, what you want to achieve. Like actually like sharing the knowledge how they use these dashboards. 26:00 Uh, have some enablement sessions, but again, relying more on, uh, you know, sharing the knowledge through champions and trying to embrace this culture within every team. 26:09 But, uh, but you know, if you just run some training for entire organization, would it work? Well, in most cases, maybe not. But if you try to have some, um, kind of 26:23 a, a bundle, uh, try to organize something like a bundle, a series of trainings, then some champions who highlight some best practices within the teams, who can explain new engineers that, "Hey, here are the places where you need to come and have a look. 26:37 Uh, here are the most common patterns which we need to have a look for our team." 26:41 Because some teams might run on, you know, like something like, um, more focused on things like, I don't know, Kafka, some messaging- Mm... streaming. Some teams are more storage oriented. 26:52 For them it would be more important to keep an eye on the DynamoDB databases and things like this. Um, so every team is unique, you know. Uh, what I learned is there is no kind of one size fits all. 27:05 Uh, so here it's more like h- having, um, organizational alignment and cultural shift. You know, make, make sure that it's clear for everyone that it's priority. 27:13 Having the network of your champions, um, and, and the, like, collection of success stories. 27:20 Uh, enablement, like some videos, some, you know, like, I don't know, office hours, like FinOps, uh, uh, FinOps office hour, office hours that people can come and, and ask some questions. 27:31 And, uh, things like newsletters. Uh, so a lot, like whatever imagination will, will, will drive you. But, um, uh... 27:39 And in all of this, you need to highlight what has been done, how it was discovered, which tools were used, and basically how you came to that sequence. 27:50 And after that, you can start looking in also in things like automation. I mean, maybe in parallel with this, so not necessarily after. But think, think also about automating something of what has been done. Mm-hmm. 28:01 Maybe also preventing some stuff to, to happen. Maybe setting up some, um, organizational SCPs, tech po- tech policies, and stuff like that. Uh, yeah. 28:13 And the last part here, what is very important is, is, is also, I think it, it, everything most likely wouldn't work, um, if there is no, um, account- accountability, uh, from- Mm-hmm... from the engineering teams. 28:27 So they should be really accountable for the cost effectiveness, cost efficiency of the applications. They should own the cost. 28:32 You know, like I really like comparison with, uh, DevOps, where, you know, we, we, we call DevOps, you build it, you run it. Yeah. 28:39 So, um, with FinOps, you, you build it, you run it, but you're also responsible for, mm- Yeah... cost effec- effectiveness, yeah, or for the cost of this. 28:48 So it adds more responsibilities and accountabilities, but at the same time, without this, it would be, it would be hard to drive the change. Yeah. 28:56 So you, you really need to push this to, uh, to the people who, who, who drive the change, who, who actually can trigger the cost impact, yeah. 29:04 [chuckles] Who actually, uh, every day deploy workloads, make decisions which actually influence the cost. 29:13 No, I, I totally agree that, especially the comparison with DevOps and security, like you have a, a, your ownership of your security failures, even though you have, you could have like a central security team that helps you prevent you making these mistakes. 29:28 Like for example, in this case, the FinOps team, uh, then you are responsible for your, the security of your, uh, tool, so you can like dive deep and see like if this cost, this is, if you are making cost enablement- Yeah... 29:41 or if the architecture is, is efficient on cost level or not. Same with DevOps. Uh, like, like as you mentioned, you are, 29:48 you own the, the time that you take to deploy something or how much it fails or how much it is available. So you should own the, the cost part. 29:58 And, and once we focus more on the ownership, which I think we are, uh, moving on since the costs are being higher, so now it's a, a problem that before it wasn't, uh, I think we will get more results and more FinOps like advanced practices. 30:13 And we are seeing that inIn businesses that are more cost, let's say, aware, such as, for example, a-airlines which play on, uh, very low margin roles that they still, for example, the, the Ryanair case which they play a lot with the spot instances to be, like, on the thin margins they have. 30:31 I assume supermarkets, uh, all the food industry will, will see, apart from banking, of course, but this is all for a general perspective. Yeah. 30:40 But banking is-- it's, it's also that the margins are not that high, so you need to play, uh, on the thin, on the thin line. Yeah. And you have to be aware of, of costs. Yeah, yeah, absolutely. 30:51 I really like your comparison of, uh, FinOps and security. I also like... 30:55 I, I think there are a lot of, a lot of parallels what we can learn from security domain, um, in FinOps as well, especially also introducing things like FinOps reviews. 31:04 We, uh, we're very familiar with the concept of, you know, AppSec reviews, AppSec approvals before- Mm-hmm... going into production. 31:12 Like, here we can reuse something like that, so have, have something like FinOps reviews, you know, making sure that, um, there is someone who familiar with, you know, like, these best practices of, um, cost of architectures within the organization actually can, uh, c- We'll have a look, uh, on your architecture before you start, you know, deploying it, before you start actually relying on it heavily in your organization. 31:36 So it's also one of the ways how you can embrace, uh, cost of architecture. Yeah. Yeah, indeed. Yeah, and- Damián? Yeah. No, it's, uh... 31:48 I also like to compare between, uh, because people often feel, feel the same also about security, about FinOps. You know, it's like people, [laughs] developers are like, "Leave me alone. Let me do my work." 31:59 But no, you need to take care also for security, you know, take care also about the cost because eventually, uh, well, we, we know what happen if you don't. Um, 32:09 and the word Focus comes, comes a lot of offense, and we see that now also from cost intelligent dashboards, we, we have not only, uh, a dashboard for AWS, but also we have for Azure and for GCP. 32:24 And actually you can do-- You can make a joint, right, uh, dashboard- Yeah... in which you show all this data. Um- Yeah... so this is very good. 32:35 And what I was wondering was, like, when can we expect to see or what are your thoughts about having a CUDOS dashboard ba-based on Focus information? Well, it's, um, it's, it's, it's a good question. 32:53 So we started, as you said, with separate dashboards for other cloud providers because there was a ask from, from customers. They wanted to have, um, to use one BI tool, you know, to, to see the data in one single place. 33:10 Um, and we also learned that, you know, building 33:15 as an insightful, uh, and as detailed dashboards, uh, for multiple, uh, cloud providers which have their own structures of billing, um, data, it's, um, it's quite tough and quite challenging. 33:29 So, um, that's why when Focus project appeared, um, I was super excited and was, you know, joining the early Focus calls and, you know, I was very excited about the project because this actually simplifies a lot for, uh, FinOps community. 33:49 Uh, so when you have normalized data across multiple cloud providers, then we as AWS, we, uh, let's say, officially joined the, uh, FinOps Foundation. We joined, uh, Focus project as well. 34:03 And, um, after that we, we contributed to the Focus, Focus project, um, as maintainers and as contributors. And we also have, um, we support Focus. We deliver data in Focus format. Um, 34:19 together with the data, what we, we couldn't, [laughs] we couldn't, uh, not to deliver a dashboard, you know, to help people not just get the data, but also visualize it and get sense out of the data. So... 34:35 And the next, and, 34:38 and the next step which we took is we now updated our GCP and Azure solutions to allow, um, actually practitioners and customers to bring data from other cloud providers in Focus format to AWS and join them in this, in single Focus dashboard. 34:57 Because one of the challenges with, um, reporting is not only normalizing the data, but also all the ETLs, you know, getting data in one place. It's, it's- Mm-hmm... as challenging as normalization. 35:10 So what-- W-we want to simplify it a bit, so we now provide solutions how to get all the data in one place, how to plug it in to Focus dashboard. 35:20 Now, as for amount of insights and details which you can get, so at the moment, you know, now, now, uh, we support Focus 1.0 in GA. Now specificati-specification itself is evolving. 35:36 1.1 version was announced and, uh, I see a lot of, you know, improvements and, um, there is a quite exciting roadmap, uh, in Focus, uh, project as well. 35:50 So with maturity of specification, we will be able to deliver more insights, more details, more reporting, more, um, kind of actionable things which customers can, can use. 36:04 So the way I see it is with maturity, uh, increase of maturity of Focus with, you know, all the feedback and all the new things which appear in Focus, it would allow us to build something very, veryClose to what we build on top of custom usage report. 36:20 Um, but obviously it took us, you know, to like four and a half years to build Kudos. [laughs] Um, so [chuckles] 36:28 I'm not saying that it will take us four and a half years to build something, uh, something like that based on Focus, but at the same time, it's also quite, quite an effort because behind every single visual there is a story, there is details how... 36:41 Like, what you can do with this because we try to also to have this kind of so what question to like, to, to pass so what question for every vi-visual which we build. Like what you can do when you see it. 36:52 Like, is it actionable at all? So, um, yeah, we... As I see it, w- I think the Focus dashboard will be evolving. We will, we will be adding more features to it. 37:05 Um, but I see it more like a parallel growth of Focus specification and then appearing in new features and capabilities on, uh, our dashboard and overall like, uh, capabilities of 37:17 anyone who use Focus, uh, regardless of the tool, you know, the like either querying grow data or any BI tool as well. 37:24 So, mm, it's quite interesting journey, uh, journey, and I hear a lot of positive feedback, uh, from, uh, customers about Focus. Um, for now, 37:35 I usually see at the moment Focus dashboard, it's more like high level, uh, dashboards which you can use to, uh, get, uh, more like general understanding of your total, you know, cost of, uh, kind of to- cost of ownership ar-around multiple cloud providers. 37:53 But then if you want to learn about like depth and breadth of optimizations, what you can improve, what the actionable insights, you still need to go, for example, to Kudos for AWS. 38:04 Um, but yeah, we will be there, [notification sound] uh, in some time. Wow. I, I wonder how it will look, uh, in the next few years when- Yeah. We'll see... 38:16 when, uh, Focus will be evolved and, uh, [chuckles] new GraphFa are gonna be out there. Um, wow. Uh, I'm, [laughs] you know, I'm a big fan and I'm looking forward to see what, what, what's coming. 38:30 Also from, uh, from graph perspective, it's the QuickSight will be also evolved? This is something that I- Yeah. 38:39 Actually QuickSight evolving go like com-comparing QuickSight at the moment with where it, it, it was like four and a half years ago when we started, it's like, it's way, way, way more feature set product, let's say. 38:57 And, um, there is, there is much way more capabilities now, uh, from both like API perspective, what you can automate on top of this, what you can use for embedding, what you can use for, um... 39:11 And also up to, uh, visual capabilities and features. Um, so now, uh, before re:Invent, QuickSight launched this high charts, a type of visual. 39:23 So it's, it kind of brings the f-full variety of completely new type of visualizations, um, where you can use this high charts kind of, uh, 39:39 visual definition within the QuickSight and have absolutely mind-blowing, you know, like very rich visualizations, which were not available in QuickSight before. 39:49 We started to do something like that in support cases dashboard, which we released, um, uh, recently before re:Invent. Uh, we will invest more in this imp- like, 40:01 uh, face lifting of, of the visuals based on this new capability in, uh, our existing dashboards as well. So we are very excited about this. And another 40:12 quite significant feature which was added, uh, by QuickSight team is ability to copy visuals between the dashboards. 40:20 And why it's important, because when you deploy your dashboard and then you want to customize it, like a lot of customers love this capability to, you know, reshuffle widgets, maybe add something which they like. 40:32 Uh, but the challenge was that when you customize your dashboard, uh, you... And we update the, uh, the new version of our, uh, provided by us dashboard template, and you update your 40:46 default dashboard, but there is, there was no way to bring this new features which we bring in the default dashboard to your customized one. 40:54 So it was a challenge for customers how then we can get this new great widgets to our customized dashboard and- Right... 41:02 and now QuickSight supports the copy, so you literally can go and copy whatever you like from the new version to your custom dashboard, place them in, in the place where you see it maybe better than we and, um, without any significant effort. 41:18 So it's... And, you know, like there is a very nice and, um, interesting roadmap in QuickSight for 2025. A lot of generic capabilities added to QuickSight, ability to integrate with Amazon Q and all the things. 41:33 So it's, it's a lot of, um, interesting, uh, features ahead, and I think it will, uh, help us also to, you know, to shape, uh, the way, um, customers approach, um, the data, you know. 41:47 Because now with all this, uh, generic capabilities, I think there will be more, uh, movement into, you know, like not getting the visualis- visualizations answering your questions, but trying to get basically your questions answered by, um, LLMs, uh, with having access to your data. 42:12 So overall, I think it will be, um, quite a shift in that direction.No, that's, that's great. 42:21 I, I mean, I think I also agree that that would be probably what it's evolving now, like being able to process the, the info to ingest it to the GenAI, and the GenAI being able to provide, like, additional information and solving questions that you may, may have already in place. 42:38 Yeah, it's, it's going to be crazy, I think, in, in 2025, the, the amount of, uh, you know, things that we are go- the amount of things that we are going to see and integrations we are going to see within the, within the AI, like now that it's more stabilized, and we see that it's more stable now, what we are going to be able to do with, with that. 42:57 [chuckles] Yeah. Yeah, yeah. But, uh, you know, for FinOps community also, GenAI, it brings quite interesting challenging in, um... 43:06 challenges, not just, uh, how to use it for FinOps, but also how to track the, you know, the cost of GenAI services, how to optimize them. 43:13 Because, uh, there is a lot, a lot of interesting new concepts appear which, uh, were not very, how to say, very common if you look at traditional applications for if- Yeah... 43:26 compute storage, uh, some, you know, databases. Uh, here it's a slightly different shift of, uh, paradigm. Yeah, especially with the, the tokenization and- Exactly... 43:39 and all the, the generation that you are not sure, like, how to, how to control it, uh, because it's, you know, something that is generative. It's very different- Mm-hmm... from, from what we are used to. Yeah. 43:49 Should we ask shorter an- uh, shorter questions, you know? And... [laughs] Yeah, it prompts like something like that. [laughs] Be concise. 43:58 [laughs] And be concise, and, and it, it always e-extends more some more, and, and it doesn't say anything after all. Mm-hmm. Uh, but yeah, uh, well, I think it's, uh, a great point to, to close up. 44:09 Uh, I think everyone, uh, knows, uh, where, where we can find you. We will put your... the links to, to your LinkedIn and to the, the dashboards, uh, GitHub, and all the information that is available there. 44:22 So, um, it's been a pleasure. I think there is a lot of valuable information that you have, uh, dropped today, and looking forward to your team, uh, developments in, in 2025. 44:34 And yeah, uh, we'll be ready to your, to your posts to, to know every, every new update in, in Cudos. Yeah. Thank you for having me. It's always pleasure. 44:44 And yeah, uh, I just wanted to say thank you again to all of our customers and all of our team. Uh, it's just, you know, being part of this is just a absolutely amazing journey. Uh, thanks. Thank you very much, Yuriy. 44:58 It was a pleasure and very enrichful, and, uh, hope that, uh, in the future we, we host you again and, uh, hear more news. Absolutely. Yeah. Thanks everyone, and see you in the next episode. Bye-bye. Bye. 45:15 [outro jingle]