Trivy and Open Source Communities with Anaïs Urlichs

Episode Summary

Corey interviews Anaïs Urlichs, an open source developer advocate at Aqua Security with a unique background - she’s never had to work with AWS. Anaïs explains how this is possible in the world of cloud and her career path from developer advocacy to computer engineering. Corey and Anaïs chat about Trivy, Aqua Security’s main open source project, an all-in-one cloud-native security scanner, and how it differentiates itself from similar offerings. Anaïs walks us through how to solve for security vulnerability fatigue with Trivy, her ultimate career journey back to developer advocacy, and concludes by explaining how Trivy is leading her to AWS after all.

Episode Show Notes & Transcript

About Anaïs

Anaïs is a Developer Advocate at Aqua Security, where she contributes to Aqua’s cloud native open source projects. When she is not advocating DevOps best practices, she runs her own YouTube Channel centered around cloud native technologies. Before joining Aqua, Anais worked as SRE at Civo, a cloud native service provider, where she helped enhance the infrastructure for hundreds of tenant clusters. As CNCF ambassador of the year 2021, her passion lies in making tools and platforms more accessible to developers and community members.

Links Referenced:


Announcer: Hello, and welcome to Screaming in the Cloud with your host, Chief Cloud Economist at The Duckbill Group, Corey Quinn. This weekly show features conversations with people doing interesting work in the world of cloud, thoughtful commentary on the state of the technical world, and ridiculous titles for which Corey refuses to apologize. This is Screaming in the Cloud.

Corey: This episode is sponsored in part by our friends at AWS AppConfig. Engineers love to solve, and occasionally create, problems. But not when it’s an on-call fire-drill at 4 in the morning. Software problems should drive innovation and collaboration, NOT stress, and sleeplessness, and threats of violence. That’s why so many developers are realizing the value of AWS AppConfig Feature Flags. Feature Flags let developers push code to production, but hide that that feature from customers so that the developers can release their feature when it’s ready. This practice allows for safe, fast, and convenient software development. You can seamlessly incorporate AppConfig Feature Flags into your AWS or cloud environment and ship your Features with excitement, not trepidation and fear. To get started, go to That’s

Corey: This episode is sponsored in part by Honeycomb. When production is running slow, it’s hard to know where problems originate. Is it your application code, users, or the underlying systems? I’ve got five bucks on DNS, personally. Why scroll through endless dashboards while dealing with alert floods, going from tool to tool to tool that you employ, guessing at which puzzle pieces matter? Context switching and tool sprawl are slowly killing both your team and your business. You should care more about one of those than the other; which one is up to you. Drop the separate pillars and enter a world of getting one unified understanding of the one thing driving your business: production. With Honeycomb, you guess less and know more. Try it for free at Observability: it’s more than just hipster monitoring.

Corey: Welcome to Screaming in the Cloud. I’m Corey Quinn. Every once in a while, when I start trying to find guests to chat with me and basically suffer my various slings and arrows on this show, I encounter something that I’ve never really had the opportunity to explore further. And today’s guest leads me in just such a direction. Anaïs is an open-source developer advocate at Aqua Security, and when I was asking her whether or not she wanted to talk about various topics, one of the first thing she said was, “Don’t ask me much about AWS because I’ve never used it,” which, oh my God. Anaïs, thank you for joining me. You must be so very happy never to have dealt with the morass of AWS.

Anaïs: [laugh]. Yes, I’m trying my best to stay away from it. [laugh].

Corey: Back when I got into the cloud space, for lack of a better term, AWS was sort of really the only game in town unless you wanted to start really squinting hard at what you define cloud as. I mean yes, I could have gone into Salesforce or something, but I was already sad and angry all the time. These days, you can very much go all in-on cloud. In fact, you were a CNCF ambassador, if I’m not mistaken. So, you absolutely are in the infrastructure cloud space, but you haven’t dealt with AWS. That is just an interesting path. Have you found others who have gone down that same road, or are you sort of the first of a new breed?

Anaïs: I think to find others who are in a similar position or have a similar experience, as you do, you first have to talk about your experience, and this is the first time, or maybe the second, that I’m openly [laugh] saying it on something that will be posted live, like, to the internet. Before I, like, I tried to stay away from mentioning it at all, do the best that I can because I’m at this point where I’m so far into my cloud-native Kubernetes journey that I feel like I should have had to deal with AWS by now, and I just didn’t. And I’m doing my best and I’m very successful in avoiding it. [laugh]. So, that’s where I am. Yeah.

Corey: We’re sort of on opposite sides of a particular fence because I spend entirely too much time being angry at AWS, but I’ve never really touched Kubernetes and anger. I mean, I see it in a lot of my customer accounts and I get annoyed at its data transfer bills and other things that it causes in an economic sense, but as far as the care and feeding of a production cluster, back in my SRE days, I had very old-school architectures. It’s, “Oh, this is an ancient system, just like grandma used to make,” where we had the entire web tier, then a job applic—or application server tier, and then a database at the end, and everyone knew where everything was. And then containers came out of nowhere, and it seemed like okay, this solves a bunch of problems and introduces a whole bunch more. How do I orchestrate them? How do I ensure that they’re healthy?

And then ah, Kubernetes was the answer. And for a while, it seemed like no matter what the problem was, Kubernetes was going to be the answer because people were evangelizing it pretty hard. And now I see it almost everywhere that I turn. What’s your journey been like? How did you get into the weeds of, “You know what I want to do when I grow up? That’s right. I want to work on container orchestration systems.” I have a five-year-old. She has never once said that because I don’t abuse my children by making them learn how clouds work. How did you wind up doing what you do?

Anaïs: It’s funny that you mention that. So, I’m actually of the generation of engineers who doesn’t know anything else but Kubernetes. So, when you mentioned that you used to use something before, I don’t really know what that looks like. I know that you can still deploy systems without Kubernetes, but I have no idea how. My journey into the cloud-native space started out of frustration from the previous industry that I was working at.

So, I was working for several years as developer advocate in the open-source blockchain cryptocurrency space and it’s highly similar to all of the cliches that you hear online and across the news. And out of this frustration, [laugh] I was looking at alternatives. One of them was either going into game development, into the gaming industry, or the cloud-native space and infrastructure development and deployment. And yeah, that’s where I ended up. So, at the end of 2020, I joined a startup in the cloud-native space and started my social media journey.

Corey: One of the things that I found that Kubernetes solved for—and to be clear, Kubernetes really came into its own after I was doing a lot more advisory work and a lot more consulting style activity rather than running my own environments, but there’s an entire universe of problems that the modern day engineer never has to think about due to, partially cloud and also Kubernetes as well, which is the idea of hardware or node failure. I’ve had middle of the night driving across Los Angeles in a panic getting to the data center because the disk array on the primary database had degraded because the drive failed. That doesn’t happen anymore. And clouds have mostly solved that. It’s okay, drives fail, but yeah, that’s the problem for some people who live in Virginia or Oregon. I don’t have to think about it myself.

But you do have to worry about instances failing; what if the primary database instance dies? Well, when everything lives in a container then that container gets moved around in the stateless way between things, well great, you really only have to care instead about okay, what if all of my instances die? Or, what if my code is really crappy? To which my question is generally, what do you mean, ‘if?’ All of us write crappy code.

That’s the nature of the universe. We open-source only the small subset that we are not actively humiliated by, which is, in a lot of ways, what you’re focusing on now, over at Aqua Sec, you are an advocate for open-source. One of the most notable projects that come out of that is Trivy, if I’m pronouncing that correctly.

Anaïs: Yeah, that’s correct. Yeah. So, Trivy is our main open-source project. It’s an all-in-one cloud-native security scanner. And it’s actually—it’s focused on misconfiguration issues, so it can help you to build more robust infrastructure definitions and configurations.

So ideally, a lot of the things that you just mentioned won’t happen, but it obviously, highly depends on so many different factors in the cloud-native space. But definitely misconfigurations of one of those areas that can easily go wrong. And also, not just that you have data might cease to exist, but the worst thing or, like, as bad might be that it’s completely exposed online. And they are databases of different exposures where you can see all the kinds of data of information from just health data to dating apps, just being online available because the IP address is not protected, right? Things like that. [laugh].

Corey: We all get those emails that start with, “Your security is very important to us,” and I know just based on that opening to an email, that the rest of that email is going to explain how security was not very important to you folks. And it’s the apology, “Oops, we have messed up,” email. Now, the whole world of automated security scanners is… well, it’s crowded. There are a number of different services out there that the cloud providers themselves offer a bunch of these, a whole bunch of scareware vendors at the security conferences do as well. Taking a quick glance at Trivy, one of the problems I see with it, from a cloud provider perspective, is that I see nothing that it does that winds up costing extra money on your cloud bill that you then have to pay to the cloud provider, so maybe they’ll put a pull request in for that one of these days. But my sarcasm aside, what is it that differentiates Trivy from a bunch of other offerings in various spaces?

Anaïs: So, there are multiple factors. If we’re looking from an enterprise perspective, you could be using one of the in-house scanners from any of the cloud providers available, depending which you’re using. The thing is, they are not generally going to be the ones who have a dedicated research team that provides the updates based on the vulnerabilities they find across the space. So, with an open-source security scanner or from a dedicated company, you will likely have more up-to-date information in your scans. Also, lots of different companies, they’re using Trivy under the hood ultimately, or for their own scans.

I can link a few where you can also find them in a Trivy repository. But ultimately, a lot of companies rely on Trivy and other open-source security scanners under the hood because they are from dedicated companies. Now, the other part to Trivy and why you might want to consider using Trivy is that in larger teams, you will have different people dealing with different components of your infrastructure, of your deployments, and you could end up having to use multiple different security scanners for all your different components from your container images that you’re using, whether or not they are secure, whether or not they’re following best practices that you defined to your infrastructure-as-code configurations, to you’re running deployments inside of your cluster, for instance. So, each of those different stages across your lifecycle, from development to runtime, will maybe either need different security scanners, or you could use one security scanner that does it all. So, you could have in a team more knowledge sharing, you could have dedicated people who know how to use the tool and who can help out across a team across the lifecycle, and similar. So, that’s one of the components that you might want to consider.

Another thing is how mature is a tool, right? A lot of cloud providers, what they end up doing is they provide you with a solution, but it’s nice to decoupled from anything else that you’re using. And especially in the cloud-native space, you’re heavily reliant on open-source tools, such as for your observability stack, right? Coming from Site Reliability Engineering also myself, I love using metrics and Grafana. And for me, if anything open-source from Loki to accessing my logs, to Grafana to dashboards, and all their integrations.

I love that and I want to use the same tools that I’m using for everything else, also for my security tools. I don’t want to have the metrics for my security tools visualized in a different solution to my reliability metrics for my application, right? Because that ultimately makes it more difficult to correlate metrics. So, those are, like, some of the factors that you might want to consider when you’re choosing a security scanner.

Corey: When you talk about thinking about this, from the perspective of an SRE is—I mean, this is definitely an artifact of where you come from and how you approach this space. Because in my world, when you have ten web servers, five application servers, and two database servers and you wind up with a problem in production, how do you fix this? Oh, it’s easy. You log into one of those nodes and poke around and start doing diagnostics in production. In a containerized world, you generally can’t do that, or there’s a problem on a container, and by the time you’re aware of that, that container hasn’t existed for 20 minutes.

So, how do you wind up figuring out what happens? And instrumenting for telemetry and metrics and observability, particularly at scale becomes way more important than it ever was, for me. I mean, my version of monitoring was always Nagios, which was the original Call of Duty that wakes you up at two in the morning when the hard drive fails. The world has thankfully moved beyond that and a bunch of ways. But it’s not first nature for me. It’s always, “Oh, yeah, that’s right. We have a whole telemetry solution where I can go digging into.” My first attempt is always, oh, how do I get into this thing and poke it with a stick? Sometimes that’s helpful, but for modern applications, it really feels like it’s not.

Anaïs: Totally. When we’re moving to an infrastructure to an environment where we can deploy multiple times a day, right, and update our application multiple times a day, multiple times a day, we can introduce new security issues or other things can go wrong, right? So, I want to see—as much as I want to see all of the other failures, I want to see any security-related issues that might be deployed alongside those updates at the same frequency, right?

Corey: The problem that I see across all this stuff, though, is there are a bunch of tools out there that people install, but then don’t configure because, “Oh, well, I bought the tool. The end.” I mean, I think it was reported almost ten years ago or so on the big Target breach that they wound up installing some tool—I want to say FireEye, but please don’t quote me on that—and it wound up firing off a whole bunch of alerts, and they figured was just noise, so they turned it all off. And it turned out no, no, this was an actual breach in progress. But people are so used to all the alarms screaming at them, that they don’t dig into this.

I mean, one of the original security scanners was Nessus. And I seen a lot of Nessus reports because for a long time, what a lot of crappy consultancies would do is they would white-label the output of whatever it was that Nessus said and deliver that in as the report. So, you’d wind up with 700 pages of quote-unquote, “Security issues.” And you’d have to flip through to figure out that, ah, this supports a somewhat old SSL negotiation protocol, and you’re focusing on that instead of the oh, and by the way, the primary database doesn’t have a password set. Like, it winds up just obscuring it because there is so much. How does Trivy approach avoiding the information overload problem?

Anaïs: That’s a great question because everybody’s complaining about vulnerability fatigue, of them, for the first time, scanning their container images and workloads and seeing maybe even hundreds of vulnerabilities. And one of the things that can be done to counteract that right from the beginning is investing your time into looking at the different flags and configurations that you can do before actually deploying Trivy to, for example, your cluster. That’s one part of it. The other part is I mentioned earlier, you would use a security scan at different parts of your deployment. So, it’s really about integrating scanning not just once you—like, in your production environment, once you’ve deployed everything, but using it already before and empowering engineers to actually use it on their machines.

Now, they can either decide to do it or not; it’s not part of most people’s job to do security scanning, but as you move along, the more you do, the more you can reduce the noise and then ultimately, when you deploy Trivy, for example, inside of your cluster, you can do a lot of configuration such as scanning just for critical vulnerabilities, only scanning for vulnerabilities that already have a fix available, and everything else should be ignored. Those are all factors and flags that you can place into Trivy, for instance, and make it easier. Now, with Trivy, you won’t have automated PRs and everything out of the box; you would have to set up the actions or, like, the ways to mitigate those vulnerabilities manually by yourself with tools, as well as integrating Trivy with your existing stack, and similar. But then obviously, if you want to have something more automated, if you want to have something that does more for you in the background, that’s when you want to use to an enterprise solution and shift to something like Aqua Security Enterprise Platform that actually provides you with the automated way of mitigating vulnerabilities where you don’t have to know much about it and it just gives you the solution and provides you with a PR with the updates that you need in your infrastructure-as-code configurations to mitigate the vulnerability [unintelligible 00:15:52]?

Corey: I think that’s probably a very fair answer because let’s be serious when you’re running a bank or someone for whom security matters—and yes, yes, I know, security should matter for everyone, but let’s be serious, I care a little bit less about the security impact of, for example, I don’t know, my Twitter for Pets nonsense, than I do a dating site where people are not out about their orientation or whatnot. Like, there is a world of difference between the security concerns there. “Oh, no, you might be able to shitpost as me if you compromise my Twitter client that I put out there for folks to use.” Okay, great. That is not the end of the world compared to other stuff.

By the time you’re talking about things that are critically important, yeah, you want to spend money on this, and you want to have an actual full-on security team. But open-source tools like this are terrific for folks who are just getting started or they’re building something for fun themselves and as it turns out, don’t have a full security budget for their weird late-night project. I think that there’s a beautiful, I guess, spectrum, as far as what level of investment you can make into security. And it’s nice to see the innovation continued happening in the space.

Anaïs: And you just mentioned that dedicated security companies, they likely have a research team that’s deploying honeypots and seeing what happens to them, right? Like, how are attackers using different vulnerabilities and misconfigurations and what can be done to mitigate them. And that ultimately translates into the configurations of the open-source tool as well. So, if you’re using, for instance, a security scanner that doesn’t have an enterprise company with a research team behind it, then you might have different input into the data of that security scanner than if you do, right? So, these are, like, additional considerations that you might want to take when choosing a scanner. And also that obviously depends on what scanning you want to do, on the size of your company, and similar, right?

Corey: This episode is sponsored in part by our friend EnterpriseDB. EnterpriseDB has been powering enterprise applications with PostgreSQL for 15 years. And now EnterpriseDB has you covered wherever you deploy PostgreSQL on-premises, private cloud, and they just announced a fully-managed service on AWS and Azure called BigAnimal, all one word. Don’t leave managing your database to your cloud vendor because they’re too busy launching another half-dozen managed databases to focus on any one of them that they didn’t build themselves. Instead, work with the experts over at EnterpriseDB. They can save you time and money, they can even help you migrate legacy applications—including Oracle—to the cloud. To learn more, try BigAnimal for free. Go to, and tell them Corey sent you.

Corey: Something that I do find fairly interesting is that you started off, as you say, doing DevRel in the open-source blockchain world, then you went to work as an SRE, and then went back to doing DevRel-style work. What got you into SRE and what got you out of SRE, other than the obvious having worked in SRE myself and being unhappy all the time? I kid, but what was it that got you into that space and then out of it?

Anaïs: Yeah. Yeah, but no, it’s a great question. And it’s, I guess, also was shaped my perspective on different tools and, like, the user experience of different tools. But ultimately, I first worked in the cloud-native space for an enterprise tool as developer advocate. And I did not like the experience of working for a paid solution. Doing developer advocacy for it, it felt wrong in a lot of ways. A lot of times you were required to do marketing work in those situations.

And that kind of got me out of developer advocacy into SRE work. And now I was working partially or mainly as SRE, and then on the side, I was doing some presentations in developer advocacy. However, that split didn’t quite work, either. And I realized that the value that I add to a project is really the way I convey information, which I can’t do if I’m busy fixing the infrastructure, right? I can’t convey the information of as much of how the infrastructure has been fixed as I can if I’m working with an engineering team and then doing developer advocacy, solely developer advocacy within the engineering team.

So, how I ultimately got back into developer advocacy was just simply by being reached out to by my manager at Aqua Security, and Itay telling me, him telling me that he has a role available and if I want to join his team. And it was open-source-focused. Given that I started my career for several years working in the open-source space and working with engineers, contributing to open-source tools, it was kind of what I wanted to go back to, what I really enjoy doing. And yeah, that’s how that came about [laugh].

Corey: For me, I found that I enjoy aspects of the technology part, but I find I enjoy talking to people way more. And for me, the gratifying moment that keeps me going, believe it or not, is not necessarily helping giant companies spend slightly less money on another giant company. It’s watching people suddenly understand something they didn’t before, it’s watching the light go on in their eyes. And that’s been addictive to me for a long time. I’ve also found that the best way for me to learn something is to teach someone else.

I mean, the way I learned Git was that I foolishly wound up proposing a talk, “Terrible Ideas in Git”—we’ll teach it by counterexample—four months before the talk. And they accepted it, and crap, I’d better learn enough get to give this talk effectively. I don’t recommend this because if you miss the deadline, I checked, they will not move the conference for you. But there really is something to be said for watching someone learn something by way of teaching it to them.

Anaïs: It’s actually a common strategy for a lot of developer advocates of making up a talk and then waiting whether or not it will get accepted. [laugh] and once it gets accepted, that’s when you start learning the tool and trying to figure it out. Now, it’s not a good strategy, obviously, to do that because people can easily tell that you just did that for a conference. And—

Corey: Sounds to me, like, you need to get better at bluffing. I kid.

Anaïs: [laugh].

Corey: I kid. Don’t bluff your way through conference talks as a general rule. It tends not to go well. [laugh].

Anaïs: No. It’s a bad idea. It’s a really bad idea. And so, I ultimately started learning the technologies or, like, the different tools and projects in the cloud-native space. And there are lots, if you look at the CNCF landscape, right? But just trying to talk myself through them on my YouTube channel. So, my early videos on my channel, it’s just very much on the go of me looking for the first time at somebody’s documentation and not making any sense out of them.

Corey: It’s surprising to me how far that gets you. I mean, I guess I’m always reminded of that Tom Hanks movie from my childhood Big where he wakes up—the kid wakes up as an adult one day, goes to work, and bluffs his way into working at a toy company. He’s in a management meeting and just they’re showing their new toy they’re going to put out there and he’s, “I don’t get it.” Everyone looks at him like how dare you say it? And, “I don’t get it. What’s fun about this?” Because he’s a kid.

And he wants to getting promoted to vice president because wow, someone pointed out the obvious thing. And so often, it feels like using a tool or a product, be it open-source or enterprise, it is clearly something different in my experience of it when I try to use this thing than the person who developed it. And very often it’s that I don’t see the same things or think of the problem space the same way that the developers did, but also very often—and I don’t mean to call anyone in particular out here—it’s a symptom of a terrible user interface or user experience.

Anaïs: What you’ve just said, a lot of times, it’s just about saying the thing that nobody that dares to say or nobody has thought of before, and that gets you obviously, easier, further [laugh] then repeating what other people have already mentioned, right? And a lot of what you see a lot of times in these—also an open-source projects, but I think more even in closed-source enterprise organizations is that people just repeat whatever everybody else is saying in the room, right? You don’t have that as much in the open-source world because you have more input or easier input in public than you do otherwise, but it still happens that I mean, people are highly similar to each other. If you’re contributing to the same project, you probably have a similar background, similar expertise, similar interests, and that will get you to think in a similar way. So, if there’s somebody like, like a high school student maybe, somebody just graduated, somebody from a completely different industry who’s looking at those tools for the first time, it’s like, “Okay, I know what I’m supposed to do, but I don’t understand why I should use this tool for that.” And just pointing that out, gets you a response, most of the time. [laugh].

Corey: I use Twitter and use YouTube. And obviously, I bias more for short, pithy comments that are dripping in sarcasm, whereas in a long-form video, you can talk a lot more about what you’re seeing. But the problem I have with bad user experience, particularly bad developer experience, is that when it happens to me—and I know at a baseline level, that I am reasonably competent in technical spaces, but when I encounter a bad interface, my immediate instinctive reaction is, “Oh, I’m dumb. And this thing is for smart people.” And that is never, ever true, except maybe with quantum computing. Great, awesome. The Hello World tutorial for that stuff is a PhD from Berkeley. Good luck if you can get into that. But here in the real world where the rest of us play, it’s just a bad developer experience, but my instinctive reaction is that there’s stuff I don’t know, and I’m not good enough to use this thing. And I get very upset about that.

Anaïs: That’s one of the things that you want to do with any technical documentation is that the first experience that anybody has, no matter the background, with your tool should be a success experience, right? Like people should look at it, use maybe one command, do one thing, one simple thing, and be like, “Yeah, this makes sense,” or, like, this was fun to do, right? Like, this first positive interaction. And it doesn’t have to be complex. And that’s what many people I think get wrong, that they try to show off how powerful a tool is, of like, oh, “My God, you can do all those things. It’s so exciting, right?” But [laugh] ultimately, if nobody can use it or if most of the people, 99% of the people who try it for the first time have a bad experience, it makes them feel uncomfortable or any negative emotion, then it’s really you’re approaching it from the wrong perspective, right?

Corey: That’s very apt. I think it’s so much of whether people stick with something long enough to learn it and find the sharp edges has to do with what their experience looks like. I mean, back when I was more or less useless when it comes to anything that looked like programming—because I was a sysadmin type—I started contributing to SaltStack. And what was amazing about that was Tom Hatch, the creator of the project had this pattern that he kept up for way too long, where whenever anyone submitted an issue, he said, “Great, well, how about you fix it?” And because we had a patch, like, “Well, I’m not good at programming.” He’s like, “That’s okay. No one is. Try it and we’ll see.”

And he accepted every patch and then immediately, you’d see another patch come in ten minutes later that fixed the problems in your patch. But it was the most welcoming and encouraging experience, and I’m not saying that’s a good workflow for an open-source maintainer, but he still remains one of the best humans I know, just from that perspective alone.

Anaïs: That’s amazing. I think it’s really about pointing out that there are different ways of doing open-source [laugh] and there is no one way to go about it. So, it’s really about—I mean, it’s about the community, ultimately. That’s what it boils down to, of you are dependent, as an open-source project, on the community, so what is the best experience that you can give them? If that’s something that you want to and can invest in, then yeah [laugh] that’s probably the best outcome for everybody.

Corey: I do have one more question, specifically around things that are more timely. Now, taking a quick look at Trivy and recent features, it seems like you’ve just now—now-ish—started supporting cloud scanning as well. Previously, it was effectively, “Oh, this scans configuration and containers. Okay, great.” Now, you’re targeting actually scanning cloud providers themselves. What does this change and what brought you to this place, as someone who very happily does not deal with AWS?

Anaïs: Yeah, totally. So, I just started using AWS, specifically to showcase this feature. So, if you look at the Aqua Open Source YouTube channel, you will find several tutorials that show you how to use that feature, among others.

Now, what I mentioned earlier in the podcast already is that Trivy is really versatile, it allows you to scan different aspects of your stack at different stages of your development lifecycle. And that’s made possible because Trivy is ultimately using different open-source projects under the hood. For example, if you want to scan your infrastructure-as-code misconfigurations, it’s using a tool called tfsec, specifically for Terraform. And then other tools for other scanning, for other security scanning. Now, we have—or had; it’s going to be probably deprecated—a tool called CloudSploit in the Aqua open-source project suite.

Now, that’s going to, kind of like, the functionality that CloudSploit was providing is going to get converted to become part of Trivy, so everything scanning-related is going to become part of Trivy that really, like, once you understand how Trivy works and all of the CLI commands in Trivy have exactly the same structure, it’s really easy to scan from container images to infrastructure-as-code, to generating s-bombs to scanning also now, your cloud infrastructure and Trivy can scan any of your AWS services for misconfigurations, and it’s using basically the AWS client under the hood to connect with the services of everything you have set up there, and then give you the list of misconfigurations. And once it has done the scan, you can then drill down further into the different aspects of your misconfigurations without performing the entire scan again, since you likely have lots and lots of resources, so you wouldn’t want to scan them every time again, right, when you perform the scan. So, once something has been scanned, Trivy will know whether the resource changed or not, it won’t scan it again. That’s the same way that in-classes scanning works right now. Once a container image has been scanned for vulnerabilities, it won’t scan the same container image again because that would just waste time. [laugh]. So yeah, do check it out. It’s our most recent feature, and it’s going to come out also to the other cloud providers out there. But we’re starting with AWS and this kind of forced me to finally [laugh] look at it for the sake of it. But I’m not going to be happy. [laugh].

Corey: No, I don’t think anyone is. It’s every time I see on a resume that someone says, “Oh, I’m an expert in AWS,” it’s, “No you’re not.” They have 400-some-odd services now. We have crossed the point long ago, where I can very convincingly talk about AWS services that do not exist to Amazonians and not get called out for it because who in the world knows what they run? And half of their services sound like something I made up to be funny, but they’re very real. It’s wild to me that it is a sprawling as it is and apparently continues to work as a viable business.

But no one knows all of it and everyone feels confused, lost, and overwhelmed every time they look at the AWS console. This has been my entire career in life for the last six years, and I still feel that way. So, I’m sure everyone else does, too.

Anaïs: And this is how misconfigurations happen, right? You’re confused about what you’re actually supposed to do and how you’re supposed to do it. And that’s, for example, with all the access rights in Google Cloud, something that I’m very familiar with, that completely overwhelms you and you get super frustrated by, and you don’t even know what you give access to. It’s like, if you’ve ever had to configure Discord user roles, it’s a similar disaster. You will not know which user has access to which. They kind of changed it and try to improve it over the past year, but it’s a similar issue that you face in cloud providers, just on a much larger-scale, not just on one chat channel. [laugh]. So.

Corey: I think that is probably a fair place to leave it. I really want to thank you for spending as much time with me as you have talking about the trials and travails of, well, this industry, for lack of a better term. If people want to learn more, where’s the best place to find you?

Anaïs: So, I have a weekly DevOps newsletter on my blog, which is anaisurl—like, how you spell U-R-L—and then dot com. That’s where I have all the links to my different channels, to all of the resources that are published where you can find out more as well. So, that’s probably the best place. Yeah.

Corey: And we will, of course, put a link to that in the show notes. I really want to thank you for being as generous with your time as you have been. Thank you.

Anaïs: Thank you for having me. It was great.

Corey: Anaïs, open-source developer advocate at Aqua Security. I’m Cloud Economist Corey Quinn, and this is Screaming in the Cloud. If you’ve enjoyed this podcast, please leave a five-star review on your podcast platform of choice, whereas if you’ve hated this podcast, please leave a five-star review on your podcast platform of choice along with an angry, insulting comment that I will never see because it’s buried under a whole bunch of minor or false-positive vulnerability reports.

Corey: If your AWS bill keeps rising and your blood pressure is doing the same, then you need The Duckbill Group. We help companies fix their AWS bill by making it smaller and less horrifying. The Duckbill Group works for you, not AWS. We tailor recommendations to your business and we get to the point. Visit to get started.

Announcer: This has been a HumblePod production. Stay humble.
Newsletter Footer

Get the Newsletter

Reach over 30,000 discerning engineers, managers, enthusiasts who actually care about the state of Amazon’s cloud ecosystems.

"*" indicates required fields

This field is for validation purposes and should be left unchanged.
Sponsor Icon Footer

Sponsor an Episode

Get your message in front of people who care enough to keep current about the cloud phenomenon and its business impacts.