Video: Protecting Azure DevOps and GitHub Data with Rubrik | Duration: 2064s | Summary: Protecting Azure DevOps and GitHub Data with Rubrik | Chapters: Welcome and Introduction (18.735s), Introducing DevOps Protection (95.245s), Cloud Management Challenges (231.865s), Automating DevOps Platforms (359.36s), DevOps Data Protection (656.25s), DevOps Data Protection (961.935s), Q&A: Azure DevOps (1178.8s), Backup and Recovery Process (1272.595s)
Transcript for "Protecting Azure DevOps and GitHub Data with Rubrik":
Hello. Good afternoon, everyone. Welcome to the session to, listen some workshop for DevOps. Today, we are joined, I'm joined with, Mike Preston, staff technical manager at Rubrik. I'm Francesco Tickham based in Amsterdam, and I'm part of the EMEA Rubrik marketing team. Today's demo workshop will give you an understanding of the security gaps, in the DevOps ecosystem, what are the critical data that must be protected in Azure DevOps and GitHub, and how our solution works for backup, recovery, and operational readiness. This is a platform that we are using where you can see different type of functions. So you have a Q and A, like a chat, and it's also possible to interact with us. So we welcome interaction. We are gonna take your questions during throughout the session. So the session is like lasts for one hour. And then I will be here to assist after a short presentation, and you'll get access to the labs. And then we'll be here to assist in case you have additional questions or want to chat with us. So thank you everyone for making time today, and off to you, Mike. Perfect. Thank you so much, Francesco. And, again, welcome, everybody. Really happy to be here. I've done a number of these workshops, in sort of the time zone for North America, but this is my first EMEA one. So this is super exciting. I got to get up a little bit earlier and, you know, get up before the rest of my family and and come down here and and deliver this. So as Francesca said, I'm Mike Preston. I work in technical marketing here at Rubrik. I've been here for, you know, almost eight years now. So quite a while, but I started we backed up VMware, and we archived that to AWS, and that was all we did. So I've seen the company grow quite a bit. I've seen the product grow quite a bit. We obviously do a lot more than that now, and and what I'm here to talk about today is really all about, sort of backing up and restoring and protecting your DevOps data. So your Azure DevOps projects, organizations, repositories, and your GitHub data and and repositories. Now as Francesco said, I do love to keep these, you know, sort of a two way street. So if you have any questions at all, feel free to use the chat, use the q and a. I can see it on my screen over here. So if you see me glancing this way a bit, you know, it's just to see if we have anything coming in. Because, you know, go ahead and and interrupt the presentation. We can steer this whatever way you wanna steer it. I gotta go through a little bit of a presentation. I promise there's not too much marketing stuff at the the the start. When we get to the end, we'll sort of get into the nitty gritty of of how all of this technically works. And then I'll paste the the lab links into the the chat box there, and you guys can go ahead and get some sort of hands on experience and see it for yourself in action. So that's kind of the agenda that we have. So, again, if you have any questions, it doesn't matter what they are. Doesn't even have to be DevOps related. You could talk we could talk about m three sixty five, although I might not have the answer. But we could talk about cloud if you want to or or anything like that. We could steer this conversation however you wanna go. So if you have any questions about Rubrik at all, feel free to throw them into the the q and a. So to kinda back up and sort of level set, you know, as to where we are in terms of IT today, clouds really change things. Right? We can see on the screen here, we have Azure, and we have things like IaaS and platform as a service, things like databases running on Azure VMs, managed databases like Azure SQL, things like AKS, Kubernetes, all of that stuff. It's a it's really changed the game in terms of how fast we could deploy things and how fast we can kinda go to market or get products built or get services running, things like that, has really added a lot of efficiency. And AI is even accelerating this even further for us. And as we in IT, we we have to manage this. Right? And it's not it's not bad for one cloud, but when you add in, you know, the the big three clouds, maybe you're an organization that uses Azure and AWS or maybe use GCP and Azure or you can you can throw OCI in there as well, it starts to get a little hectic in terms of management. Right? And that's not to mention we probably all have things like enterprise data centers running things like, hypervisors, like VMware or Nutanix, things like that, SaaS applications like m three sixty five, Salesforce, those big, you know, unstructured data, whether they're, you know, things like object storage in the cloud or, you know, those big giant mass systems that we have, sitting in our in our data centers today, as well as things like our identity systems. So active directory, Entre ID. Right? These turn into a full time job to manage as well. So all of a sudden, we've we've grown from just having to manage data centers to having to manage three different clouds, data centers, SaaS applications that that just seem to to keep coming as well as giant unstructured data and identity. And in order to be successful as in IT to manage all this because this is this is crazy. Like, this is hard, and I don't have to tell you guys that. You're you guys are out there living this now. We have to manage this. We have to make sure we're backing all this up. We have to secure all this. And the only way to really do that successfully is to start automating and building different scripts and and automations around that. Right? It's the only way we can actually do it because we we only have so much time in one day to to do all this stuff. And if we wanna do this and do it successfully, you know, we really have to build automations around that. And when I say automations, you know, it's a can be kind of boiled down into three main things, like application deployments, updates, security fixes, as well as things like infrastructure management. So think like infrastructure as codes, things like Terraform and and, you know, Ansible in a way. Like, basically, these plans that we can have to manage the creation, deployment, and ongoing maintenance and configuration of our environment. So we build out all of these great automations, and quite often, we store them in some sort of DevOps platform. Right? That's where we're gonna store the source code for these things because we wanna have versioning on it. Because now all of a sudden, these automations are, you know, super important because they're the source of truth for our environment. So we leverage things like GitHub, like GitLab, like Azure DevOps to store all of this this data for us. And, you know, it it works great. We have things like CICD pipelines that run from our DevOps platform and connect into our clouds, into our data centers. And we could start to do things like continuous deployment, do things like continuous integration, do things like, you know, security testing and different stuff within the pipeline to ensure that we're actually deploying, what we wanna deploy within our environments. However, we do a great job in terms of, when we think about, you know, protection or data protection. We do a great job at protecting what's in the middle here. Right? Our Azure VMs or AWS EC two instances, our object storage, Kubernetes, RDS, Azure SQL. Right? We do a great job at ensuring that we have backups for that. We understand we need backups for that. However, this DevOps platform, which is sort of the source of truth, tend to get sort of a little bit ignored. And the problem is when it goes down, you know, if it's compromised, all of a sudden, we can't do all those application updates or those security fixes or deploy those those infrastructure as code plans. Right? So without those platforms, without all of that data in the repository, without our pipeline configuration and all of that, you know, we're kinda left in this limbo state where our source of truth is gone, and then we have to manually manage this environment again. And and we all know that's not good. So, you know, the quest sorry. I should say, like, it's not, you know, just me saying this. Right? We're seeing this more and more, I would say, ever since around 2020, things like supply chain attacks, you know, things people trying to disrupt the GitHubs of the world or the GitLabs or the code builds or or anything like that to try and disrupt that to ensure that you can't deploy those pipelines. You can't run those pipelines. You can't get access to those codes. Right? Maybe it's people just changing a package name slightly in some sort of open source project. It gets accepted, merged, and all of a sudden, you know, we have compromised code being deployed throughout the world. And we've seen tons of different examples of this. You know, one that's kinda recent, which sorta caught my eye because I thought it was a little bit unique, was Wiz kind of talked about a compromise that they saw within AWS CodeBuild and and GitHub. And the way it worked was CodeBuild actually, you know, did a check. It looked at the GitHub user that was trying to execute the pipeline, looked at that user ID, and sort of looked to see if it had a match of the list of trusted IDs that it had. And the problem here was it was one character in in that regex string that that where it looked for that match, and that's the little tilde or the hat. I'm not sure what it's called. A shift six on your keyboard, that little arrow hat thing. Basically, what that does is say, hey. You need to be an exact match, and that wasn't there. So, really, it was doing, a substring match. So if you had a user ID in GitHub of, say, one two three, and that one two three was in, was anywhere within the ID of a trusted user on code build, it would match and say, hey. You're you're okay to proceed. You're a trusted user even though the exact there was no exact match. So what would happen is, you know, they could use automations to create user IDs until one matched, open a PR with that user, merge that PR. It would trigger that code build, pipeline run. Code build would look at that user ID and say, hey. Oh, this this looks like a trusted user. I think we're we're we're free to proceed here. And then, of course, they stick malicious code into that PR, which abuses build context, grab credentials to the entire GitHub, organization, And then all of a sudden, they could use those said credentials to wreak havoc, whether that's, you know, deletion, encryption, things like that. So, thankfully, this one never actually made it out into the wild. Like, Wiz had discovered it and and found it, and it all got patched up before, everything sorta had a chance to spread. But it's just an interesting concept of how one character could, like, basically, could have brought down entire GitHub orgs. You know? And it's it's it's a scary world out there, so we need to make sure we're protecting this. So that's my question, I guess, to everyone here today is what's your strategy right now to protect your DevOps data, to protect GitHub, to protect Azure DevOps? Do you have one? And it's okay if you don't because a lot of organizations really haven't thought of it. A lot of organizations are are barely scratching the surface on on protecting their actual production data that's in the data centers or in the clouds today. So it's really okay if you don't have one, but the idea is you should probably start thinking about, you know, protecting that source of truth before something bad happens because these types of attacks are just getting more and more common. Now when I go out there and I talk to people, you know, I do hear some pushback. Right? I hear, yeah, I I can protect my my DevOps data right now. I could I have things like data version history. You know, I write scripts to perform manual snapshots or create Git bundles or do the odd Git pull or Git clone every now and then. Or, you know, I don't need backups because I have a thousand developers, and they have all this stuff sort of sitting local on their backups as well. So they're syncing back and forth with the repository, so I have all these federated local copies out there. And, you know, these are these are great. You should you should definitely take advantage of these things. But in the world we live in today, there's huge gaps that that these types of, I guess, you could say, quote, unquote, backup leave. Right? So native version history is a huge cyber resilience gap. And what I mean by that is, you know, anytime we have if if you consider versions backups, then anytime you have backups that sit right alongside the same storage as production or as your source, you know, that's a huge problem. So if someone gains access to that GitHub back, organization, they're not gonna leave those versions around. Right? They're they're gonna try and purge everything they can. Right? They're not gonna be like, oh, well, I'll just leave the versions so they could restore. Right? Anytime that you have your backups can't sit right alongside your your source data. Manual snapshots is a huge data protection gap. Right? You need to ensure that this is automated or, scheduled in some sort of way. You can't depend on running that Git bundle or or Git clone before making a change every single time because someone's gonna forget, someone's gonna make a mistake, some something might be corrupt. Who knows? Right? You might not be able to restore this stuff. It's a it's a huge issue. You really need to ensure that you know, if you use some sort of enterprise grade data protection software to back up your data, you should really have that same solution to back up that source of truth and not not rely on just a manual snapshot of that. And then there's a federated local copies, which, honestly, it's a great great concept. We have data all over spread throughout. However, merging that back in can be a bit of a nightmare. Don't know if you guys have went through merge conflicts and things like that, but they can get pretty tricky. So if if you can imagine having, you know, 30 or 40 different repos coming in trying to merge into main and trying to rebuild that, you know, that'd be that'd be insane. I have a hard enough time merging one branch into another when there's complex trying to figure out what needs to be kept at and what needs to go away. So while these solutions these three solutions are definitely needed, you should definitely take advantage of them. They themselves, are not a backup solution for your data, which is why Rubrik is here, which is why we've built our DevOps protection. And you could see there's really three main three main, I guess, pillars that we we base our DevOps protection on, which is cyber resilience, automated protection and recovery, and then the unified enterprise enterprise grade solution. And what I mean by that sort of cyber resilience is, you know, you're gonna get the same immutable backups that you get with every other supportive workload in Rubrik with your DevOps data. So your GitHub, your Azure DevOps. When we perform backups of that, those backups are gonna be marked as immutable so they can't be changed. So when, you know, an attacker does gain access either into your GitHub environments, your Azure DevOps environment, into whatever, they're not gonna be able to purge those backups as well. Those backups are really gonna be there as your last line of defense. They're gonna be air gapped, you know, tucked away, things like that. So, you know, they're not living on the same storage as your actual source. In terms of flexible recovery and protection, you know, you get that same SLA based engine. I'm not sure how many people here are new to Rubrik versus how many, you know, have used Rubrik before, but, you know, our SLA based automation is or our policy based engine is essentially the concept of defining your end state. Right? I wanna have a backup every twelve hours that I keep for a day. Maybe I'll take it daily. I'll keep it for a month, so on and so forth. You define that end state. You apply that to your environment, and away you go. That same policy engine gets applied to your your Azure DevOps data, your GitHub repositories, and and whatnot. So you just simply apply an SLA to those. All of a sudden, we begin taking backups and adhere to that. In terms of recovery, we're gonna give you options to recover, you know, across different GitHub organizations or across different Azure DevOps projects or or organizations. So you have the ability to completely abandon a a, say, an Azure DevOps project or an Azure DevOps org If that identity sort of, boundary has been breached, you can completely build rebuild somewhere else if need be. And then finally, one that kinda gets skimmed over, but, you know, is is definitely important, is going to be sorry about that. It's going to be, you know, sort of the simplicity and, efficiency of the platform itself of having one unified enterprise grade solution. So this is one platform that's gonna back up your production data. It's gonna back up your identity, and it's gonna back up your DevOps data. Right? So it's one interface to learn granular RBAC, which is a huge, know, driver, especially in terms of of your GitHub data because quite often, it's not your backup administrator that's coming in. It's not the same guy that's backing up Nutanix on prem or m three sixty five that's gonna come in and manage your GitHub backups as well. So you're able to sort of grant, you know, those granular permissions to your GitHub administrator or to whomever. Maybe it's your your, your, DevOps architect or somebody that's going to take control of these backups, you can basically give them the access they need in Rubrik to get the job done, in terms of just your DevOps data. And, of course, those security features like retention, locking, quorum authorization, you know, all the the old security features that Rubrik's built into the platform apply to your DevOps data as well. So in the end, you know, really, what what we've aimed to build here is enterprise grade protection for your enterprise critical code. Right? This is this stuff's important. We need to make sure we're actually protecting it, just like we are all of our other data. So with that, let me take a quick glance at q and a because I haven't clicked on that. Oh, I do see a question there. That's great. So for Azure DevOps, Leah's a great question who says, Azure DevOps, is the entire contents of the organization backed up? Like, you know, the work items, wikis, pipelines, repos, things like that? And that's a great question. And as of, you know, version one, we focus just on the actual code, so the repository itself, as well as the pipeline. So things like work items and and boards and wikis and all of that is a fast follow coming in version two, and you'll see that coming out very, very soon, like, within, I believe, it's within the quarter. I can't commit to any times, but it's definitely short term road map. So, for people that don't know, you know, this this sort of DevOps backup, with Rubrik was was integrated into our system through an acquisition of a company called BackRightUp who did support things like wikis and work items and things like that. So as we we sort of port that code over into Rubrik Security Cloud, to give you that unified platform, you'll see that come out. But that's a great question. Same goes for GitHub. At this moment with version one, it's, repositories and pipelines or GitHub actions, which is basically the YAML file sitting inside that, that repo. So great question there. Thank you, Lee. Okay. So in terms of how this stuff works. So I told you I I'm trying it through the marketing stuff as quick as I can. I tried to make it not too dry, but this is gonna get even more dry unless you're a technical person. But this this it's always great to know sort of how all this works under the hood and, you know, what we can do in terms of that. So let's talk a little bit about I'll talk a little bit about onboarding for Azure DevOps and a little bit about onboarding for GitHub. But then when I talk about the backup and recovery pieces, sort of how that all works, I'm going to sort of just do Azure DevOps because the backup process, the recovery process, pretty same for for ADO and GitHub. So in terms of onboarding, how's this all gonna work? Well, the first thing we're gonna do is we're gonna ask for some credentials into your Azure DevOps environment. So the customer's gonna provide that, enter those into Rubrik Security Cloud, and then we're gonna go ahead and create the resources that we need in terms of creating a service principle and application resin registration and whatnot. And that's what's gonna be used for subsequent access into the environment. We're also gonna create two permission groups within ADO. So one's gonna be for backup, and one's gonna be for restore. So that what this allows us to do is, you know, customers don't need to give us long standing write permissions into their environment to take backups. Instead, they could just give us those sort of read only permissions that we need to take backup. And then when we do need to perform a restore, they can elevate that, so we can, you know, pick up those those permissions when we do need to perform a restore and then take them away after if you do not want Rubrik having, you know, long standing right permissions to do restores. So that's handled through permission groups. Those permission groups, basically, that service principal gets assigned to to be a member as those. And then, of course, we we spark up our exo compute within the environment to do things like all of the data movement, the git pull, the git clone, all those sort of commands that need to be executed. So that's onboarding for Azure DevOps. GitHub, very, very similar. Again, right, we're gonna provide those OAuth credentials into GitHub. We don't have the concept of permission groups in GitHub. So instead, we have these little applications that get deployed. Same deal. One for backup, one for recovery. We don't need those long standing write permissions. You can just give us the permissions required for backup, only and then elevate the restore. We're going to attach sort of that organization. I I didn't have this on the Azure DevOps slide, but also the Azure DevOps gets attached to some sort of of object storage. So that's where the backups are going to live. And then, of course, we deploy that exo compute again, and, you know, metadata is sent back to RSC, customer data, DevOps data, that all lives under your own cloud account, under the your own control of of the customer itself. Now in terms of how production works. So here we have, essentially, you know, a couple different tenants. One one that's storing our Azure DevOps org. We have another one over here where we're sort of logically air gapping our our data into an Azure storage account. So how the actual backup works is the control plane here, Rubrik Security Cloud, that's delivered via a SaaS based interface. It's gonna send the request for exit compute to be instantiated. We're gonna spark up sort of those temporary pods. If it's the first full backup that we're running, we're gonna run a git bundle. We're gonna generate a bunch of metadata, and we're, of course, gonna copy that over to the storage account where we mark it as immutable and and all that fun stuff. So that's the first full backup. Subsequent backups are all going to be incrementals. Right? So in order to perform an incremental, we do need to reference that latest full. So we're gonna pull that latest full back down into our temporary storage. We're gonna go ahead and run a git pull against the the repo and compare it to that latest full to sorta get the latest commits and changes, as a base. So that gives us a little incremental backup. And then from there, we're gonna bundle that, write that to storage, and away we go. That gets marked as immutable and everything like that. And then, of course, everything gets spun back down, and, the, metadata gets sent back up to Google Security Cloud so we could do things like, you know, fast search and and whatnot. But that's essentially how the backup process itself works. I'm gonna take a pause here because I do see some questions about licensing. So licensing, let's see. Mark has a question. What's required to back up Azure DevOps? There is, this is its own, skew within Rubrik, so definitely talk to your sales team there. You know, there's obviously different volume discounts, things like that, but it is priced, I believe, per repository that you're looking to protect. Now would be a great time to talk to those sales teams because it is priced very, very aggressively at the moment because, again, it's it's relatively new within Rubrik, released within the last three months. So that's it. Another question here from Tar is asking about AKS. He's like, can it be private and use only private endpoints? It absolutely can. So we do support what's called private exocompute. I should mention too that this you know, within the next three to four months as well, we're we're actually offering a fully hosted version of of this as well where you don't even need to have Kubernetes running in your environment. Rubrik will have Kubernetes running within the Rubrik environment, and Rubrik will handle, you know, all of this storage as well. So sort of a fully hosted, fully managed solution by us. You just simply apply SLAs to your your DevOps org or your project or your repos or GitHub org or your GitHub repos, and, you know, the storage, compute is all handled by Rubrik. So moving on to recovery, sort of how that works. So now we have our full backups, a bunch of incremental backups sitting tucked away. Again, we're gonna send that request to instantiate exo compute. We're gonna download all the necessary bundles that we need, to basically get to that point in time that you've selected. So you can almost think of this almost like a database and transaction logs. Right? It's pretty similar. So we download a full and a a bunch of incrementals to get there. We're gonna reconstruct that repository, you know, to that desired point in time that you've selected. We're gonna run a git verify against that to ensure that we can actually recover this stuff, ensure that we have data integrity and things like that. And then from there, we're gonna take that repo, push that back up into the a new repository. So for those familiar again with Rubrik terminology, we usually have in place recovery and export in place overwrites the original. Export creates a new one for DevOps, for both Azure DevOps and GitHub. We're only supporting export, and this is intentional through conversations with some different customers that we had on the beta that said, you know, we don't wanna disturb any existing merges or any existing pipeline runs or anything that might be running on the source repository itself. So we're always going to restore into a new repository, perform that cleanup, and away we go. All done. So with that, it's kinda your guys' turn now. Take a quick look at the questions here. See anything else. Oh, I see. How about LFS files? Are those backed up? I'm not sure what LFS files are. I'll try and get an answer for you, Pippa, around that, and, I'll let I'll I'll put it back in the chat for you, in terms of what that is. But I'll I'll dig into that, but let me get the labs unleashed for everybody here so we can all do do our thing. Okay. So, before I jump in, two labs, I do wanna sort of, you know, give you a bit of a tour of them because there are some, you know, some tricks and different things that you can, you do in the labs. So they're they're all broken down. We're gonna have one for GitHub, one for Azure DevOps. Feel free to do them both. Feel free to just do one. It's up to you. They're gonna be broken down into chapters. Right? So we have things like onboarding, how to do how to configure our back, how to back up, and and essentially how to restore. So you can jump around from those different chapters. You can click around as it prompts you to click on different things, or you can just use keyboard arrows as well if you want to. It helps you navigate a little bit better. But, yeah, those are sort of my tips and tricks for for, you know, navigating around in some of these hands on labs. I'm gonna go ahead and drop these labs into the chat at the moment. So we're gonna have two there, with GitHub and with Azure DevOps. So feel free, and I think Francesco, and I will be here for a little bit if you have any questions. And with that, that's basically what I had for the presentation for you today. Francesco, anything else to Absolutely. No. Thank you so much, Mike. Like, a super like, a a very, like, a very detailed presentation and, like, very straightforward as well. So everyone just gives us signal if you are able to access the labs. If there is any question, please feel free to drop it in the chat or in the q and a as you are doing already, so we are able we'll be able to monitor it. So the session will go ahead, up until 2PM CET. So if you're in The UK, it will be, like, 1PM. If you are elsewhere, like, it depends on your on your time zone, but we still have other thirty minutes. I guess the labs will be open for at least additional twenty four hours. So in case you want to revisit any of the steps, please feel free to do so. We are also able to I'm also dropping my email in the chat. So in case you need to be connected with anyone here at Rubrik for additional information to have a conversation about our solution and you want to dig down on what is the like, the your real application case. Here is my email, so I can direct you to the right people if you don't already have a contact with your rubric. But with with this, like, thank you so much. Like, me and Mike will be in the background of four day, coming, half an hour. So, enjoy the lab, and, let us know if you have questions. Thank you. Thanks so much. Hello, everyone. We are coming to the conclusion of today's session. Thank you again, Mike, for the presentation, for guiding us through the labs, and everyone for making time to attend today. As said, live the labs will be still available for additional hours. So in case you have to revisit any step, feel free to to keep playing with the with the environment. If you have questions, feel free to contact us. I dropped my email here. I'm more than happy to put you in touch with our team in case in case you need any type of of clarification. And thank you again, Mike, for participating. Thank you, everyone. And please check our calendar of workshops and camp rule breaker. We have a full full calendar through to the month of April. So we are gonna give you additional opportunities to test the Rubric Security Cloud with additional pieces of our solution. Thank you again, and I hope you enjoy a great day and a good weekend, which is approaching faster. For the majority of us, that probably is a short week. So thank you again, everyone. Have a good day.