Automating DevOps Workflows with Generative Artificial Intelligence – Asaf Yigal, Logz.io

This is texturong TV. Hey guys. Thanks to the throw.

We're here with esophage go who's CTO for logsy Io and we're talking about generative Ai and how it will be applied to devops a soft. Welcome to show. Yeah, thanks for thanks for having me.

This has got to be one of the hottest topics in recent memories. So what's your thoughts here is there's some way to kind of use this within the platforms. We have and embed an extended and to what extent will it automate a whole bunch of things that today or essentially manual efforts that kind of drive us all crazy.

Yeah, that's a good question. So definitely there is room for it. And we've just integrated the check GPT with with logsdale.

I think the biggest challenge of devopsies is the mounting data that is just keeps on growing and growing and in the ability to make sense out of it. So companies connect more and more data to help them troubleshoot to help them understand what's going on in their environment the more data they collect the harder it is for them to understand what is actually going on the way we've integrated Church GPT, which is this is the first time and and we're gonna continue to involve more capabilities of GPT into into the DeVos, but the way we did it is we integrated with our cognitive insights. So our cognitive Insight is a patented technology that scans the internet for known issues in it and in devops, so we collect links and signature and if we find it in your logs that we highlight it what we did now is we kind of like Should GPT to ask if this happens to you, what would be the most likely scenario?

What would be the most likely cause of this problem? And we just paste it kind of like that the resolution and We were surprised at the level of intelligence and the level of insight it can bring in how much more information it can assist with with devops. And with the operators were trying to address no issues and unknown issues.

So we do see it as a way to enrich the data with insights from The Wider community and to reduce significantly reduce the time it takes to troubleshoot today. We've also seen these platforms right some code. Quality of that code can be questionable.

But what's your sense of how integrated will all these platforms get within our devops workflows. I mean, you know, we're going to use them to write scripts and are we going to use them to automate additional processes or how far will they go? I think yes definitely into devil space and into the development space.

There's going to be some. Integration of church GPT on writing code and writing scripts and connecting between systems. I can tell you that.

For example, we have an Integrations team. So when we don't know how to do it when we don't know how to integration works if this is an obscure device that we don't know. Today they go to action for them is to us check Deputy how to integrate these two systems together and it gives them really it gives them a solution that they haven't thought of before it's not writing the code.

It's not implementing the connectivity but it just gives them the concept in the idea which is taking from the community. So definitely that I think there is still there's still kind of like a compliance and and illegal question whether you can actually write code with it and integrated with your production with with when it comes to to specific intellectual property because it's really unclear who actually wrote the code. It's not like the Church of these writing the code is just taking it from somewhere else.

So that's going to be once that issue is going to be addressed and definitely it's going to be used wider for writing codes and writing tests and and doing stuff like that. One of the issues that everybody in devops deals with is burnout. There's just a lot of manual effort and a lot of stress and strain on the systems these days.

Do you think that that's gonna be reduced as we rely more on these types of platforms to automate some of these tasks? I think so. Yes.

I think some of it is a lot of the stress out and the burnout that happens the days because the level of stress these guys are under and because the motive different dimensions of different system that they have to know how to maintain and how to operate and the ability to acquire that knowledge in real time when there is an issue in the system when things can when people cannot log in where people cannot buy in your e-comers, there's a lot of pressure coming out on on these people and they definitely spend days and night being awake troubleshooting stuff. I think the ability to help them troubleshoot is it has the potential of alleviating some of that pressure and especially for junior people until they become more expert. It's definitely going to help with alleviating some of the pressure from them.

How smart will all this get because one of the challenges is training the language models to stay current and the IT environments are changing almost, you know a couple of times a day now people are dropping new code in the systems. So can these things keep up with that rate of change? I think I think so.

I think that a lot of companies today. They utilize a lot of code that wasn't written yesterday. A lot of the whether to use is even if they upgrade the version every hour every two hour the integrate open source, that may have been written a month ago two months ago two years ago five years ago and and this information is included today.

Development is and and for a lot of companies a lot of e-commerce and gaming companies and the like is leveraging existing engine leveraging existing open source and connecting them in the most efficient way in the most unique way and it's not about generating new code. That is that is solving a problem. If this is this is usually the core IP of every company is usually a relatively small in kind of like what they do and a lot of the code that actually breaks stuff that they in right because if they wrote it then it's easier for them to troubleshoot it easier to make sense out of it and if they didn't write it if someone else wrote it as an open source, this is where it gets complicated because they don't know how it runs and when it breaks, it's hard to troubleshoot what's going on?

One will the ultimate impact of all this be on devops careers because it seems like a lot of the entry level tasks may be automated. So we just gonna have people working at higher levels of abstraction, maybe a few of them or you know for that matter or people gonna be eliminated all together. I think we see it with every every new technology that's being included that people's the job gets easier, but it creates more opportunities at the end of the day.

So if you see for example Cloud adoption and the docker and kubernetes adoption, you would think that people their life would be easier because now there is an abstraction layer that can do it. But in reality you can do more and you can do more than people actually do more and there's more demand for for people more demand for companies. If it's going to be easier for a company to get off the ground running with an existing product.

They didn't have to wait one two, three months like they do today it can they can do it in one two, three days. You would see more companies you would see more technology. You would see more more people being demanded for it.

So. I don't think that this and the mundane tasks of figuring out how an open source operates and what happens when it breaks which today the information exists and devops people today when something breaks and something that they don't know what they do is they Google it up and they search the internet and they read 10 different posts that tell them what the problem is. They choose one that seems like it's a likely solution to try to implement it.

This is how they do their investigation. No one's going in and reading the actual code so all it's going to do is reduce the time to resolution which is great A lot of these people are better better life balance and faster troubleshooting for issues that today they don't know how to solve. I don't think that this is going to impact the kind of implement in this so fewer people will be able to manage higher levels of it at scale.

I guess the other second question is that well more people adopt devops best practices because the barrier for acquiring those skills is going to be a lot lower. possibly it is possible that the level is going to be Cumberland become lower it is possible and and you would see kind of like if people try to do more with less which I think is the Kind of like the error that we live in but they still want to do more. We're still in deficit when it comes to the amount of work needed to be done in the technology space with companies every company that we talked to including ourselves as well.

We wish we had more people to do the work. We wish we we could generate more work from the existing people that that we have. Um, and if that's gonna help that technology is going to help us generate more with the same amount of people that we have.

Then this is going to be Trend in the industry itself. Do you think there will be one or two of these types of platforms out there are ultimately will there be I don't know hundreds of them all tuned for specific use cases that are better than each other but they're all generative AI platforms to one degree or another. It's a good question.

I have no idea. I think there is definitely potential for specialization there. So if it's going to be one that is like a medical oriented and whether it's going to be devops and ideally into the one that's like travel oriented if I want to book a flight search for the simple one instead of me going through a bunch of sites.

I don't know kind of I don't have any any prediction on this as well. I think that if something is is definitely providing a solution that it's going to attract people and you see it with the church GPT that it's they they really came out of out of the blue and they just had a great product and this is how it's creating the trend and how everybody's running the trend in Google and Microsoft and everybody is is behind this generative AI So once you're best advice to folks because a lot of people tend to get nervous when you know, we start inventing technologies that might move their proverbial cheese. How do we kind of think about this today?

And you know, what should people be doing? So first of all, we cannot we cannot stop change. We know it for the whole story of hundreds of years that when things change you have to adopt it and you have to make the best out of it.

I think we need to kind of like join it figure out. What's the best way to leverage that technology if it has any negative impacts on things and and continue to adopt it. All right, folks you heard it here change is coming whether you like it or not.

So you might as well get it because otherwise it's just gonna run you over so there you have it the South thanks for being on the show. Yeah. Thank you very much.

All right back to you guys. Thanks.