E04: Clean Up Your Infrastructure to Cut Idle Costs
Nick and Jason discuss the basics of infrastructure cleanup and its significance in cloud cost optimization. They highlight the ease of use of cloud infrastructure and the accompanying difficulty of cleanup. They dig into why companies need to understand how their cloud infrastructures work and that cleanup is never a one-and-done deal. Companies that scale will always run into waste issues so it's best to get educated on your infrastructure and cut unnecessary costs.
What's up everybody? This is Nick from tenacity cloud.com with my co-host Jason, and you're listening to the Cloud cost optimization podcast. More people than ever are building stuff in the cloud and spending a lot of money doing it. So on this show, we're gonna discuss some of the cloud cost optimization trends that are in the market, and the things that you can do to, you know, build your own cloud cost optimization strategy.
So welcome to episode four. Today we're gonna talk about sort of infrastructure and. You know, getting that right, getting configuration, right, the drift over time, what happens, what to look for in, in sort of how to approach it. In short form we'll dive into more nuance around the things you can do in the, the specific strategy, specific tactics even done in the environment operationally.
But today we'll cover kind of at a high level what you want to do from a infrastructure cleanup perspective. So, Jason you know, would love to get kind of a just a first thoughts on. Infrastructure cleanup in its importance in cloud customization. Fire away, man. [00:01:00] Well, well first may, maybe the first step is to realize that you have a bunch of waste
You know, like a lot of, a lot of companies that we talk to, and I probably experienced this as well, is It, it's happened more than once where we identify a bunch of stuff that's not being used because it's underutilized or it's not attached, or whatever the case may be. Right? And the first response is always very similar.
No, no way. Not, not, not possible, right? And, and even from, Team members that are managing set environment, and it's not really their fault. This stuff is hard to find, right? I mean, if, if I go create a thousand snapshots in EU West two, which nobody logs into as an example, right? Or something like that, if I replicate them over there.
And no one ever goes to that region with the native tools within, like say AWS or, it's hard to find those things. And so, you know, the first step step I think, is [00:02:00] just to recognize that you most likely have things out there that you're paying for, that you shouldn't be. and two, that that's completely normal because one of the benefits of cloud is that you can spin things up and spin and, and turn them down.
You can do testing, you can, there's, it's so many benefits that you wouldn't want to take that away. But one of the results, one of the byproducts of the, of how easy it is to use cloud infrastructure and services is that a lot of those things get left up and in a lot of cases they shouldn't be. . I, I think that's a really important point is you know, there's a, there's a time function here and there's a kind of an obfuscation function here because as hardware has become software it's really easy to write a piece of code to well do things repetitively, number one, but also just do, do things really, really quickly.
Sometimes just out of the purview of, of, or oversight of a single. And or single engineer, which, which, you know, we all, we all knew this person, right? Or these people in our past who kind of had a, an iron grip [00:03:00] on, you know, what was deployed where, because it was very physically you could put your eyes on it, you could touch it and they knew it cause they had to rack a stack and connect it, manage it, et cetera.
Well, we live in a very different world now and so, you know, given this time function and I. I have observed in, in our consulting, you know, when you get to a year, in two years, there seem to be these hurdles that, that it, people who've gone through transformation get a year into the new world or in twos in the new world and they start to see some really funky things in, in their environment from a spending perspective.
This is about the point when they go, Wait a minute, I thought this was supposed to be cheaper to run in the cloud. Oh my gosh, we gotta move workloads back. They start to have a little bit of a panic component to it in. You know, our, our advice by and large at that point is, hold on. You, you're just experiencing the fact that, that that things were spun up and, and left.
And, and you need to actually think about, well, what is the stuff in the environment you don't need any longer? Because you really should treat it as e femoral. You really should treat it as, as temporary infrastructure that's there for a workload there for a need needs to go away. [00:04:00] So I think that's kind of the first thing to, to, to think about.
I, I, I think, you know, the second thing to really understand here is that it's it's not hard to clean any of this stuff up. Like it's not hard to, to get rid of E two instances you don't need, It's not hard to go set your s three versioning policies correctly or to clean up snapshots that are suddenly.
Now two years worth of, of data as opposed to the, you know, maybe 30 days that you need, or six days or whatever it is. But the, the, the actual. You know, caveat to that is once you know, right, it's actually finding it that that can be the problem in, in environments at scale. And I, and I think that's what we've run into, is that people are surprised cuz it's suddenly been found.
And then the, the knee jerk reaction is, well no, that can't possibly be abandoned infrastructure. That's gotta have a need. Mm. You know, I, and, and we certainly have had experiences where there have been significant impacts to cost spending in, in our environment, right? I mean, we've, we've experienced our own environment.
Yeah. We, we [00:05:00] use our own tools. We eat our own dog food. And, and does it happen? Absolutely. Do we need to go in there? Do we need to review? What's been created, maybe even what's been destroyed, because we have teams of people working in this stuff. Absolutely. We need to do that. And so even companies that do this for a living need to do this, right?
like everyone should be taking a look at this, reviewing it on some sort of recurring basis and ensuring that you know, what their teams are doing and what they're building is you know, being. Torn down when not needed anymore. And it's not just that, I mean, it's also, you know, if you've got stuff out there that you don't know is out there that you're paying.
You're most likely not securing those resources in a way. You know, you're, not only are you wasting money, but you're also probably leaving some holes open within your environment and your network for bad actors to take advantage of. And you know, I know this. It's not necessarily cost [00:06:00] related, but that will end up causing you, that will end that could potentially impact you from a spend perspective if you have to notify people if say, data has been breached.
I mean, I think the one, the one that sticks in my head is like maybe a few years ago, Capital One left an S3 bucket. It was either an S3 bucket or an easy two instance that they didn't know they had that had access to set S3 bucket, which had a bunch of customers date in it, right? Not used. Should have been deprecated, but was left up.
And people got a bunch of credit card numbers out of it, right? So, and Capital One's pretty advanced when it comes to cloud usage. So even those companies that have been doing this for years and years and years and years in public cloud environments are susceptible to these things as well. In fact, the more you're used really important, the more you're susceptible then it's a really important point is, is you know, abandoned infrastructure, infrastructure.
Unmanaged is unmanaged attack surface. It's, it's, it's there. So you, you know, your cloud, your cloud spending strategy [00:07:00] is really, can be a part of your cyber security strategy as well. And, you know, you can ask a couple of old hats in cyber security. That's us. What's the first thing we do when we start thinking about.
An environment that we're consulting on or we're taking over, or we're building the operational plan for the very first thing we do is trying to minimize how big it is, right? We wanna, we wanna mm-hmm. , make it as small as possible. Why do you have all of this infrastructure when you're trying to do this one thing?
Like this is what we should be focused on. And so that's always the first step. So I, you know, I think it's also important to recognize that. Part of what happens here, especially to those who are new in the cloud. But I think this, this lives in people who've been in the, the industry for a long time is that the way we pay for technology is completely different now.
And so sometimes the old mentality of, you know what, I already pay for that up front. So if it sits there, I don't really care. I paid for it three years ago and, and I'm gonna use it and I'm gonna use it next month. And so it can just kind of sit there idle and, and not really do [00:08:00] anything. Well, that paradigm has shifted, right?
So when we're in public cloud you know, those workloads now cost you money while they're just sitting there idle. What you really want to do is be shutting those. Right. You wanna be getting rid of those. You wanna be standing that up for the once a month that it's needed or the once a quarter mm-hmm.
or if it's, you know, maybe, maybe even change your paradigm on, you know, how, how to invoke that workload. It doesn't need to traditionally be a server, right? Maybe we can leverage serverless infrastructure containers to, to help us with those workloads that are periodic ephemeral as opposed to the mentality of, you know, and well fall victim to this, fall victim to this of, of, you know, it used to be that I bought stuff and I set it up and.
Kinda let it go. I already paid for it, so why not let it sit there idle? It wasn't hurting you. Yeah. I mean, now it hurts you. Yeah, it does, I think. I think it hurt you back then as well. I think that is more visible by the resource. Like you, like we, we actually, you can actually see how much you pay for an individual resource.
Now that was, you couldn't do that when you had, you know, a bunch [00:09:00] of VMware, you know, server farms, whatever you wanna call them in your, in your data center. Right? So I, I think the effects were still there because at the end of the day, if I'm using more. For things that I don't need, I'm probably gonna have to buy hardware sooner.
But it was, you know, when you, when you're making that decision today, you might not feel the effects of it for two years or a year, year and a half, whatever the case may be. And then it just becomes part of the business's normal buying cycle. Like now you're, now you are just estimating as the IT manager.
You're, you're basically just including the waste in your, in your budgets for the next year and how much hardware you need to buy. Well, now that you can get down to the individual resource, it, it is cheaper. It's way cheaper. You just have to know what you're doing. You have to make sure you manage the environment right.
Yeah, absolutely. And, and I think that that you know, Jason and I have discussed this a lot especially with, with our clients and, and with our customers is, you know, when, you know how, how to invoke sort of this [00:10:00] cleanup strategy in your broader. Your your broader, broad, broader cost optimization strategy.
And it's really important to do this first. It's really important to at least understand it first before you get into a commitment based discount strategy, because you don't wanna be committing to spend or committing to use for workloads that are unnecessary. So that's, that's number one. So at least having your arms around it, understanding what you need to exclude and so on, so that, Buy the right commitment based discounts have the, have an appropriate strategy there is really important.
It's also really important before you get into the refactor stage because the refactor stage of cloud cost optimization, you don't wanna be refactoring a workload that you'll find out later is like, really not being used as intended or shouldn't be there in the first place. Or, you know, may maybe, you know, you don't understand the refactor it needs to go through because you don't understand kind of the use pattern or.
It's just configured incorrectly, right? So, really important to kinda get your arms around this first. And it's ongoing. It's not just a, it's not just a one and done. This is, this has a time function to it. So, I can't, [00:11:00] you know, I can describe actually the number of times that, that we've gone into consult and.
You know, went on retainer to help an organization advise, be part of their cloud, center of excellence and watch over time as we come in and initially do a cleanup and then see that decay, that drift, that happens over time, where the next quarter we're already looking at workloads that we're part of a project that, that we're pointing out and going like, Well, hey, didn't this project end last month?
Why is this still here? You know, there's an expiration even set on it, like, why is it still here? And it, and so just, you know, having. Become a, a part of the ongoing process. I, you know, I don't know if there's anything else. Do you have any parting words Jason, before we close out the episode? Yeah. Stop, , stop letting your cloud provider rip you off.
Turn off your
They're not ripping you off, but yeah, it's designed to be, You should not be buying, We don't need to buy Jeff Bezos a second support yacht for his yacht. Right. We, let's, let's. Let's try to fix that [00:12:00] hashtag Jeff Bezos or something like that. , right? We, we laugh about this from a security perspective of like, infrastructure does not become, does not come secure by default for the reason of, they want it to be really easy to use and, and part of that ease of use is so you'll continue to spend money there.
I mean it. It is, it is brilliant and it is true about product across the spectrum outside of aws, Azure, anywhere in the world. Make it easy, easy to use, people will, will consume it. Cloud usage and cloud spending is cloud friendly, not business friendly. It's, it just, but, but it's the nature of their business, right?
They're giving you the power to pay for things by the sip. You're gonna have to sift through the mess. Absolutely. Well, I think that's a really good note to end on. So, thank you everyone for tuning in to the Cloud Cost Optimization Podcast. To next time we'll dive another one of these topics.
Thanks so much.