Listener Questions 5
Links: Cloud FinOps: https://www.amazon.com/Cloud-FinOps-Collaborative-Real-Time-Management/dp/1492054623 FinOps Foundation: https://www.Finops.org/ AWS cost management blog: https://aws.amazon.com/blogs/aws-cost-management/ Mastering AWS Cost Optimization: https://www.amazon.com/Mastering-AWS-Cost-Optimization-operational/dp/965572803X TranscriptCorey: This episode is sponsored in part by LaunchDarkly. Take a look at what it takes to get your code into production. I’m going to just guess that it’s awful because it’s always awful. No one loves their deployment process. What if launching new features didn’t require you to do a full-on code and possibly infrastructure deploy? What if you could test on a small subset of users and then roll it back immediately if results aren’t what you expect? LaunchDarkly does exactly this. To learn more, visit launchdarkly.com and tell them Corey sent you, and watch for the wince.Pete: Hello, and welcome to the AWS Morning Brief: Fridays From the Field. I am Pete Cheslock.Jesse: I’m Jesse DeRose.Pete: Wow, we’re back again. And guess what? We have even more questions. I am… I am… I don’t even know. I have so many emotions right now that are conflicting between a pandemic and non-pandemic that I just—I’m just so happy. I’m just so happy that you listen, all of you out there, all you wonderful humans out there are listening. But more importantly, you are going into lastweekinaws.com/QA and you’re sending us some really great questions.Jesse: Yeah.Pete: And we’re going to answer some more questions today. We’re having so much fun with this, that we’re just going to keep the good times rolling. So, if you also want to keep these good times rolling, send us your questions, and we’ll just—yeah, we’ll just roll with it. Right, Jesse?Jesse: Absolutely. We’re happy to answer more questions on air, happy to let you pick our brains.Pete: All right. Well, we got a couple more questions. Let’s kick it off, Jesse.Jesse: Yeah. So, the first question today is from Barry. Thank you, Barry. “New friend of the pod here.” Always happy to have friends of the pod. Although I do feel like that starts to get, like, Children of the Corn, kind of. I think we started that, and I also am excited about it, and also upset with myself for starting that.Pete: That’s all right. Friend of the pod. Friend of the pod.Jesse: “New friend of the pod here. I work in strategic sourcing and procurement and I was curious if there are any ways that you recommend to get up to speed with managing cloud spend. This is usually closely monitored by finance or different groups in product, but I can see a significant potential value for a sourcing professional to help, also.” And that’s from Barry, thank you, Barry.Pete: Well, I’m struggling not to laugh. “This is usually closely monitored by finance or different groups in product.”Jesse: Yeah…Pete: But I mean, let’s be honest, it’s not monitored by anyone. It’s just running up a meter in a taxi going 100 miles an hour.Jesse: Yeah, that’s the hardest part. I want everybody to be involved in the cloud cost management practice, but there’s that same idea of if it’s everyone’s responsibility, it’s no one’s responsibility. And so this usually ends up at a point where you’ve got the CFO walking over to the head of engineering saying, “Why did the spend go up?” And that’s never a good conversation to have.Pete: No, never a good one. Well, Barry because you’re a friend of the pod, we will answer this question for you. And honestly, I think it’s a great question, which is, we actually have been working with a lot of larger enterprises and these enterprises still have their classic sourcing and procurement teams. That’s not an expertise that is going away anytime soon, but like most teams within the company that are adopting cloud, it’s obviously going to evolve as people are moving away from, kind of, capital intensive purchases and into, honestly, more complex, multi-year OpEx style purchases, with cloud services and all the different vendors that come with it. It’s going to just get a lot harder.I mean, it’s probably already a lot harder for those types of teams. And so there’s a bunch of places I think that you can go that can help level up your skills around cloud spend. And I would say the first place that I personally got to dive in a little bit more—I mean, my history has been using Amazon cloud and being a person who cared about how much my company spent on it, but when you—joining Duckbill, you need to dive into other areas around the FinOps world. And the book, the O’Reilly book, Cloud FinOps is actually a really great resource.Yeah, I think it’s really well written and there’s a lot of great chapters within there that you can kind of pick and choose based on what you’re most interested in learning about. If you’re trying to learn more about unit economics, or you’re trying to learn more about how to monitor and track things like that, it’s a great book to dive into, and becomes a really great reference that you can leverage as you’re trying to level up this expertise within yourself or your team.Jesse: It’s a really, really great resource. The other thing to think about is any kind of collaborative social spaces where you can be with like-minded individuals who also care about cloud costs. Now, there’s a number of meetups that exist under the FinOps title that may be worth looking into. Obviously, we’re recording this during the pandemic so I don’t recommend doing those in person. But as you are able to, there may be opportunities for in-person meetups and smaller local groups focusing on cloud cost management strategies together. But also check out the FinOps Foundation. They have a Slack space that I would love to tell you more about, but unfortunately, we’re not allowed to join. So—Pete: Yep.Jesse: —I can’t really say more about it than that. I would hope that you’re allowed to join, but they have some strict guidelines. So, I mean, the worst that can happen is they say no; it’s definitely worth signing up.Pete: Yeah, and they have to us. [laugh].Jesse: Yeah.Pete: I think when you get into the FinOps Foundation, you should angrily say that we should have more FinOps experts in here like the great Jesse DeRose should be a member of this one because right now, he’s just framed his rejection notice from there, and—Jesse: Oh, yeah.Pete: —while it looks beautiful on the wall, while I’m on a Zoom with him, I want more for you, Jesse.Jesse: I want more for me, too. I’m not going to lie.Pete: So, I don’t know this might sound a little ridiculous that I’m going to say something nice about AWS, but they have a fantastic cost management blog. This is a really fantastic resource, really incredible resource, with a lot more content more recently. They seem to be doing some great work on the recruiting side and bringing on some real fantastic experts around cost management.I mean, just recently within the past few months they talk about unit economics: How to select a unit metric that might support your business, talking more about unit metrics in practice. They start at the basics, too. I mean, obviously, we deal a lot in unit economics and unit metrics; they will start you off with something very basic and say, “Well, what even is this thing?” And talk to you more about cost reporting using AWS organizations for some of this. It’s a really fantastic resource.It’s all free, too, which is—it’s weird to say that something from AWS is free. So, anytime that you can find a free resource from Amazon, I say, highly recommend it. But there are a lot of blogs on the AWS site, but again, the Cost Management Blog, great resource. I read it religiously; I think what they’re writing is some of, really, the best content on the blog in general.Jesse: There’s one other book that I want to recommend called Mastering AWS Cost Optimization and we’ll throw links to all these in the [show notes 00:07:30], but I, unfortunately, have not read this book yet, so I can’t give strong recommendations for it, but it is very similar in style and vein to the Cloud FinOps book that we just mentioned, so might be another great resource to pick up to give you some spot learning of different components of the cloud cost management workflow and style.Pete: Awesome. Yeah, definitely agree. I’d love to see, again, more content out here. There’s a lot of stuff that exists. And even A Cloud Guru has come up with cost management training sessions.Again, we’d like to see more and more of this. I’d love to see more of this come from Amazon. I’d love to see—you know, they have a certification path in all these different areas; I’d love to see more of that in the cost management world because I think it’s going to become more complex, and having that knowledge, there is so much knowledge, it’s spread so far across AWS, helping more people get up to speed on it will be just critical for businesses who want to better understand what their spend is doing. So, really great question, Barry, friend of the pod. We should get some pins for that, right? Friend of the pod pins?Jesse: Oh, yeah.Pete: And yeah, really great question. Really appreciate you sending it and hopefully that helps you. And if not, guess what? You can go to lastweekinaws.com/QA, and just ask us a follow-up question, Barry. Because you’re a friend of the pod. So, we’ll hopefully hear from you again soon.Jesse: Thanks, Barry.Pete: Thanks.Announcer: If your mean time to WTF for a security alert is more than a minute, it’s time to look at Lacework. Lacework will help you get your security act together for everything from compliance service configurations to container app relationships, all without the need for PhDs in AWS to write the rules. If you’re building a secure business on AWS with compliance requirements, you don’t really have time to choose between antivirus or firewall companies to help you secure your stack. That’s why Lacework is built from the ground up for the cloud: Low effort, high visibility, and detection. To learn more, visit lacework.com.Pete: All right, we have one more question. Jesse, what is it?Jesse: “All right, most tech execs I speak with have already chosen a destination hyperscaler of choice. They ask me to take them there. I can either print out a map they can follow, procedural style, or I can be their Uber driver. I could be declarative. I prefer the latter for flexibility reasons, but having said that, where does one actually start?Do you start with Infrastructure as a Service and some RDS to rid them of that pesky expensive Oracle bill? Do we start with a greenfield? I mean, having a massive legacy footprint, it takes a while to move things over, and integrating becomes a costly affair. There’s definitely a chicken and egg scenario here. How do I ultimately find the best path forward?” That question is from Marsellus Wallace? Thank you, Marsellus.Pete: Great question. And I’m not just saying that. I guess I have a question. Or at least, maybe we have different answers based on what this really looks like. Is this a legacy data center migration?The solution here is basically lift-and-shift. Do it quickly. And most importantly, don’t forget to refactor and clean up after you shut down your old data center. Don’t leave old technical debt behind. And, yeah, you’re going to spend a lot, you’re going to look at your bill and go, “Holy hell, what just happened here?”But it’s not going to stay that way. That’s probably—if you do it right—the highest your bill is going to be because lift-and-shift means basically just moving compute from one location to another. And if you’re—as we spoken about probably a million times, Jesse and I, if you just run everything on EC2 like a data center, it’s the most expensive way to do the cloud stuff. So, you’re going to then refactor and bring in ephemerality and tiering of data and all those fun things that we talk about. Now, is this a hybrid cloud world?That’s a little bit different because that means you’re not technically going to get rid of, maybe, physical locations or physical data centers, so where do you start? It’s my personal opinion—and Jesse has his own opinion, too, and guess what it’s our podcast and we’re going to tell it like it is.Jesse: [laugh].Pete: [laugh]. You know, my belief is, starting with storage is honestly a great way to get into cloud. Specifically S3. Maybe even your corporate file systems, using a tool like FSX. It’s honestly why many businesses start their cloud journey, by moving corporate email and file systems into the cloud.I mean, as a former Microsoft Exchange administrator, I am thoroughly happy that you don’t have to manage that, really, anymore and you can push that in the cloud. So, I think storage is honestly a great way to get started within there: Get S3 going, move your file systems in there, move your email in there if you haven’t yet. That’s a really great way to do it. Now, the next one that I would move probably just as aggressively into and, Marsellus, you mentioned it: RDS, right? “Should we move into RDS, get rid of expensive Oracle bills?”Yeah, anytime you can pay ol’ Uncle Larry less money is better in my mindset. Databases are, again, another really great way of getting into AWS. They work so well, RDS is just such a great service, but don’t forget about DMS, the database migration service. This is the most underrated cloud service that Amazon has in there, it will help you migrate your workloads into RDS, into Amazon Aurora. But one thing I do want to call out before you start migrating data in there, talk to your account manager—you have one even if you don’t think you have one—before starting anything, and have them help you identify if there are any current programs that exist to help you migrate that data in.Again, Amazon will incentivize you to do it, they will provide you credits, like map credits or other investment credits, maybe even professional services that can help you migrate this data from an on-premise Oracle into AWS, I think you will be very pleasantly surprised with how aggressive that they can be to help you get into there. The last thing that I would say is another great thing to move in our data projects. So, let’s say you want to do a greenfield one, greenfield type of project into Amazon, data projects are a really great way to move in there. I’m talking things like EMR, Databricks, Qubole, you get to take advantage of Spot Fleets with EMR, but also Databricks and Qubole can manage Spot infrastructure and really take advantage of cloud ephemerality. So if, like I said, you started by pushing all your data into S3, you’re already halfway there on a really solid data engineering project, and now you get to leverage a lot of these other ancillary services like Glue, Glue DataBrew, Athena, Redshift.I mean, once the data is in S3, you have a lot of flexibility. So, that’s my personal opinion on where to get started there. But Jesse, I know you always have a different take on these, so where do you think that they should start?Jesse: Yeah, I think all of the recommendations you just made are really, really great options. I always like to look at this from the perspective of the theory side or the strategy side. What ultimately do these tech execs want to accomplish? Is it getting out of data centers? Is it better cost visibility?Is it optimizing spend? Is it better opportunity to move fast, get new RaD things that you can’t get in a data center? What do these tech execs ultimately want to accomplish? And ask them. Start by asking them.Prioritize the work that they want to accomplish first, and work with teams to change their behaviors to accomplish their goals. One of the biggest themes that we see in the space moving from data centers into cloud providers or even just growing within a given cloud provider is cost visibility. Do teams know why their spend is what it is? Do they know why it went up or down month-over-month? Can they tell you the influences and the drivers that cause their spend to go up or down?Can they specifically call out which teams or product usage increased or decreased, and what ultimately led to your spending changing? Make sure that every team has an architecture diagram and they can explain how they use AWS, how data moves from one service to another, both within their product and to other products. Because there’s definitely going to be sharp edges with data transfer between accounts. We’ve seen this happen to a number of clients before; I’ve gotten bit by this bullet. So, talk to your teams, or talk to your tech executives and have those tech executives talk to their teams to understand what do they ultimately want to accomplish?Can they tie all of what they’re trying to accomplish back to business metrics? Maybe a spike in user logins generated more usage? If you’re a photo storage company, did a world event prompt a lot of users to upload photos prompting higher storage costs? Are you able to pull out these specific insights? That’s ultimately the big question here. Can you boil it down to a business KPI that changed, that ultimately impacted your AWS spend?Pete: I think this is a scenario of where you get started. Why not both? Just maybe do both of these things that we’re saying.Jesse: Yeah.Pete: And honestly, I think you’ll end up in a pretty great place. So, let us know how that works out, Marsellus, and thank you for the question. Again, you also can send us your questions, and we will maybe answer these on a future episode; lastweekinaws.com/QA, drop a question in there, put your name, or not or a fake name, or even a joke. That’s fine, too. I don’t know what the text limit is on the name, Jesse. Can you put a joke there? I don’t know. You know what? Test that out for us. It’s not slash QA for nothing. So, give that a little QA, or a question and answer and [unintelligible 00:17:29]. All right. Well, thanks, Jesse, for helping me out answering more questions.Jesse: Thanks, everybody for the awesome questions.Pete: If you enjoyed this podcast, please go to lastweekinaws.com/review, give it a five-star review on your podcast platform of choice, whereas if you hated this podcast, please go to lastweekinaws.com/review and give it a five-star rating on your podcast platform of choice and tell us, what would be the last thing that you would move to AWS? It’s QuickSight, isn’t it?Jesse: [laugh].Pete: Thanks, everyone. Bye-bye.Announcer: This has been a HumblePod production. Stay humble.