Integrating Databases into DevOps with DBmaestro’s Gil Nizri

This is Techron tv. Hey guys, thanks for the throw. We're here with Gil Nery, who's CEO for DB Maestro, and we're talking about, well, databases and why today they're still not quite part of the DevOps workflow to the degree they should be, and what we all should do about that.

Hey, Gil, welcome to the show. Thank you. Thank you for providing me with this great opportunity.

This has been going on for some time now, and we've always had sort of database administrators over on the left hand side somewhere, and a bunch of DevOps teams running around on the right hand side. Sometimes the DevOps teams is running the database, usually because some developer picked up something that looked like a document database. But why do we have this disconnect today and what can be done about it?

Okay, so let me start first by saying a couple of words about my company, DB Mastro. So, database Maestro is a database DevOps platform solving a huge problem in the DevOps world today, the manual database change management. So let me, Mike, let me ask you something.

Who runs the world? Like I know what Rihanna think. Rihanna think that that's girls runs the world.

But from your perspective, who runs the world? It, it all depends who I'm talking to. Most days, developers think they run the world, but then again, DevOps engineers say, Hey, you guys work for us.

I will, I will help you. Applications runs the world today. Everything we do in our life, every corporate in, in the world, doing everything he can using application.

So the ability to deliver application faster, secured compliance, consist and more efficient is what makes DevOps critical for every company. So let me take you through a scenario. Now, you are a DevOps engineer, just successfully completed a DevOps project.

So your application delivery process is fully what I call git automated. So the entire application is now really, uh, under Git. And are you safe?

Now? Think again. I don't think so because there is a significant part of your application that is neglected.

The database code that GID lives behind untouched is neglected. That doesn't make any sense to me because the database code portion from the entire application size ranges between 10 to 40% for the sake of our conversation. Let's say 20%.

So after you're so happy and completing your costly DevOps project, you ended up where your application is 80% compliance, 80% secured, 80% efficient. You got my point. It's, it's a binary thing.

You cannot be 80% compliant, so you have to have DB Mastro, and that's exactly what you're doing. Doing what we are doing. We are completing this 20% that is so critical to make DevOps complete.

So another question that you may ask is like, w why is it so complex to do it manually? Why can't I make it? Why can't my team make it manually?

And this is a conversation I run with lots of, uh, CIOs and, and, and, and decision makers. And I say, you know what? Let me walk you through a typical customer scenario.

So supposedly, Mike, you, you work for a company, well, classic enterprise company in, in, in, in, in America, and you guys have 200 applications, which is not too much. It's only 200. So each of each uses four database environments like testing development, UAD production.

If you did a multiplication, it means that you run 800 environments that you have to deal with and make sure that they are aligned sanction. And so, but let's say that you are a Superman and you can do 800 good, but these company have anything between 30 to 50,000 changes of database, database changes a year. So now you do the multiplication and you understand that it's nearly, it's not nearly, it's impossible to run this thing, uh, manually.

And if it's not complex enough, let's say that in your company you have thousand, thousand developers in five different locations and maybe three time zones. So it's a recipe for chaos. And the question whether you have, uh, errors and pitfalls, and for god's sake, downtime is not.

If I will have it is when will I have it? And people kind of ignore downtime, but downtime is a very critical thing because, you know, even through Gartner's conservative assessment, the cost of a one minute downtime is $9,000. Tell that to McDonald's.

In last March, they had a downtime of one hour calls from a software delivery, uh, for one hour, 90,000 stores of, of McDonald's didn't sell. So the damage, estimated damage not by McDonald's is a hundred million dollars for one hour. That's it.

So, so make any calculation if Amazon have that in Black Friday or something like that. So that's, that's the paradox here. Like, you know, people ignore that because they say, well, we are 99% uptime.

But you know, sometimes statistics can be bad and you are not. 01%, and when it comes, you pay cash. So let me ask you, is this a technical challenge or is it more of a cultural issue?

Because we just grew up with all these database administrators running around and they don't wanna give up what is essentially their job? Amazing question, Mike. Amazing question.

So you see, you and I, we understand the databases one of the most important corporate assets, right? It all the, the wisdom of the company reside in the databases, all the knowhow, all the skills of everything. So why steal DBAs, DevOps engineers and see those as techies and not as business stakeholders?

That's, that's actually your question, was that I will tell you why, because they find it challenging to elevate themself to that position to understand the database is not a skill statement. Database means business agility for the business. If your database is shut down, if your application is shut down, the business doesn't work, think about a bank, and now the loan application is shut down.

So the guy from the database or the guy from the DevOps now trying to figure out how do I solve this problem and go, all the teams now are hectic and trying to figure out the solution, but in the meantime, this amazing bank doesn't give a dollar in loans. So it means millions of dollars every minute, boom, boom, boom. And still we think that they are techies wrong.

Big, big mistake. They're not techies. They should be considered as business stakeholders.

Hmm. How do we incorporate databases into a DevOps workflow? And I'm asking that question because, um, I guess there's more APIs out there, but there's also more types of databases than ever.

So you know what to use when and where and how to programmatically manage it. Seems a little bit daunting. Good.

So first, as as an FYI DB MASTRO is database agnostic. It means that we give you a unified repository or a unified methodology and, and, and, and process how to, um, manage your database change process across all databases. So think about it, you have a real position of governance here.

It's not just managing the process. You govern the entire, as we agreed a minute before the entire database assets of the organization and you can deal with it through compliance and and, and security issues and others. Are people building applications that span multiple databases or are the applications themselves getting more complex or are they just kind of picking one for each application?

And it's a different use case for each one, but they manage them in semi isolation. Excellent. So like, you know, maybe 10 years ago when you were releasing a a, a new version of the application, you were already doing it maybe 12 times a year, once a once a month, once a quarter.

But as I said, applications runs the world. What does it really means when I say that? It means that you have to release versions sometimes three times a day, not once a month, three times a day.

So the ability to deliver fast and secured and compliance and uh, consistent is actually transformed into business agility. That goes back to your previous que question about is it a business or not business? It is a business because it's a, it's the new competitive advantage of organizations to be able to launch a delivery or to deliver, uh, their, their um, changes promptly.

Now, I have, uh, talked to many of our customers. One of them is CDW, you're probably aware of you, you familiar with this customer? And the guy from his name is Adam, the guy from CDW tell me, told me that they uses DB maestro to gain a competitive edge in their industry because now they can launch new versions in the speed of light and they're secured and they're fully automated.

So it's a hands free enough to the Thursday night release parties, which are not really a pizza parties. There's like, you know, a manual process where everybody is now on Zoom and he said, John, please release your piece. He's releasing it.

Good team do yours. He does. Oh, it doesn't work.

Let's roll back. Come on guys, we are in 2024. We don't do this stuff anymore.

And, and these kind of parties are no longer needed. You can do the whole thing hands free. And that's exactly what we did in for, for CDW.

Um, most of the process, if not all of it, is fully automated and hands-free. What impact do you think AI is gonna have on all this then? Because there's a lot of capabilities being tossed around from predictive degenerative.

Will AI save us from ourselves? Yeah, that's a great point. And from my background, uh, in my previous life I had a ai, uh, startup.

We kind of build a data scientist in a box, but this is a different story. So I'm kind of familiar with AI and definitely with generative ai. And we recently, like 3, 4, 5 months ago, we released the, uh, error management powered by, uh, ai.

So like now when you coding your database code and it doesn't work for, so for the reason we immediately bring you what the generative AI think that is a proper or a better, uh, uh, way to do it. So we give you a recommendation, you can take it or not, but it's definitely you have a feedback immediately. And the feedback of an ai, and this is just the tip of the iceberg because my team now is extremely busy on adding so much value around ai.

I believe in the next three to five years, uh, we are, we will change the industry using generative ai, but not only generative AI in general, um, to, to extract way more value and for, for our customers, uh, from their, uh, release data. How do I go about bringing these teams together then? Do I just put everybody in a room and lock the door till reason prevails or is there some other way of thinking about this?

Well, you go, you're just asking the million dollar question, you know, if I'll give you this answer, I have to kill you. But, but it's challenging. It seems that you really, really good, you familiar with the challenges we are facing on a day to day?

Because as, as we mentioned before, now the DBAs and the DevOps are kind of having the mandate for the database, but yes, but when you're dealing with the security of the database, you need the CSO guys and you're dealing with stuff that you need the data architects and you need, well, the challenge is always is to find, I I'm calling him the mega sponsor. You need a mega sponsor because if you will start only from bottom up, you will be challenged because the guys on the, I would say lift level, they have no ability to gather this, uh, amazing team that is so diverse team that are, most of them are not from their department by the way. They're from different departments.

So your question is good is is really good, but it's challenging. But we have, you have to figure out how you do a buy-in properly and, and it's a process. You don't sell DevOps in one week.

It doesn't happen. It's a process. It's a journey that due to the journey, you have to spread tons of value all over the place of the, the enterprise all over the place and to regain and gain more stakeholders, those who will vote for you, you don't sell to a single person.

Yeah. You make the DBAs like it as a solution. That's great.

Now you have to engage them and help them to have a buy-in in their organization. That's practically answering your question. Will the DBAs continue to manage the databases or maybe some subset of the databases while the DevOps teams take on more of the database operations?

I mean, what would be the shift in balance of roles there? Because you know, I've seen some DevOps teams where they do manage the database. Well, it's, I think it's somehow related to, uh, cultural things in organizations and the basic of cultural thing is the adherence of change.

You know, like, you know, changing stuff is not necessarily something everybody loves or I would say the other way around and, and it depends, it depends on the spirit of the commander, it depends on the sense of liability and who really owns it. Because you know, the DBAs or DevOps are the gatekeepers, right? But their boss is the one that kicks out if something bad happens, not them, they're just, you know, they are doing the job.

But, and, and it depends if you are in the C level, the guys of the C level do understand the real importance and then they could decide which, which would be the perfect, uh, uh, one to lead from. From my perspective, DBA is a DevOps guy. He just doesn't know that.

But he is DevOps guy. Mm-Hmm. He is doing, uh, something in the DevOps process.

So I think at the end they could merge together. Like DBAs will have some sort of a sub role or role in, in, in, in DevOps because it appears that DevOps is more sexy and more trendy and more hype. So I believe they will merge to somehow.

All right folks, you heard it here, that merge is all but inevitable as the application environments get more complex, we're processing data in real time and the business is more dependent on software than ever. So let's all come together 'cause it's just a question to win. Hey Gil, thanks for being on the show.

Thank you so much for, for this great opportunity. Have a good one. All right.

And back to you guys in the studio.