Shailesh Yadav

Hello. I hope you can hear me. My name is Shailesh Yadav. Um, from India originally. At the moment, living in New Zealand. And, uh, about my background, I started my, uh, IT career back in 2009. And that time I was a, um, Linux administrator, then I became a software developer. Then I moved into, uh, build and release engineering space and eventually eventually ending up into automation and DevOps around, uh, 2017. And, uh, since then, I have been working in in New Zealand. And, uh, oh, yeah. That's about it.

Uh, for, uh, for looking from the perspective of scaling, Issues related to an application. I would be, uh, looking into, uh, the logs first and, uh, looking for the hogs kind of processes, uh, that are involved into the application are on time. Uh, basically, that means, If if there any request, we're just taking too long. That sort of stuff. That will be my first, uh, kind of investigation, uh, from investigation perspective. And then I will be, uh, looking into the resource consumption of the application itself. I have mostly worked in Azure. So I'll be looking into, uh, the monitoring, Uh, you know, the the metrics for the performance metrics for the application. Uh, I think this is where I'll get started when it comes to scaling ModelX. Uh, because that will give me an idea on, uh, how this, uh, you know, application is gonna scale out and what could be the potential problems while doing that.

Uh, I think the answer, uh, is more on the technical terms for this question. And, uh, what I will be looking for, uh, from my understanding is that this is a classic case of using, application of Kubernetes, because, you know, it has, uh, tremendous, uh, uh, you know, uh, power to basically, deal with this kind of situation where we can, uh, you know, we can, uh, very easily and quickly, uh, replace our failing notes, uh, kind of, uh, you know, notes and applications, uh, in a very, uh, very easy easy way kind of.

Uh, I think we'll use for that, we'll use namespaces and, uh, and, uh, you know, names namespaces are the the logical boundaries for different different multi, uh, uh, you know, multi multiple environments. So I'll be using, uh, namespaces And, uh, the relevant, uh, state files for, uh, for, basically, uh, you know, separating, uh, the involvements. This is what, I guess

I think that we will be using the standard, uh, standard logging and metrics solutions for this kind of situation. I don't have anything on top of my head which could be, uh, kind of a game changer, but I'll go in with the standard stuff on this.

Uh, I don't have much context into what this is all about, but, uh, you know, from a very, uh, very, uh, you know, very In first looks, I would say that, uh, the one thing which I find an issue when it comes to maintainability is that, Things are hard for me. Hello? Yes. Yes. Last Yes. Last Okay. Bye. Yeah. So what I was saying was that, uh, I Find that a lot of things are hard coded here, which are not idle for from my understanding. I would just, you know, uh, parameterize all of these. That would be something really, uh, really sought after in terms of maintainability. And also with the password. Where is the password coming from? Oh, password is also a plain text, Which is, uh, no title. It should be coming from some somewhere else, like, Azure Key Vault or something. Something should be there. Sound good. So, basically, it seems like it's a connection DB connection being made up, uh, from literal strings, which is

Uh, I have not used much of AWS. Uh, actually, I've used, to be honest, in back in 2000 Sixteen something. Uh, 3 years I spent working in exclusively in AWS involvement, but never worked on GCP. But, uh, from my previous experience, uh, I would say that, uh, uh, it's a very actually, the question is very very open ended So I would say that, you know, I would be more looking into the application design itself, you know, because one of the problems I'm not sure if you know, but, I've been I've been looking into the updates and to how things are going in terms of cloud. And 1, uh, there's a strange pattern which is emerging that people are moving away from now. They are going back on premises. And one of the things is, uh, the cost. Right? So one of the things is the cost. And the reason why that there's cost is that, uh, their sources Our, uh, the applications the solutions are consuming too much of resources. And one of the reason behind that is that they are just doing a lift and shift. They are not optimizing, uh, the applications, uh, applications or the solution itself. We are just doing, uh, you know, literally lift and shift. And that's why they don't have because, you know, clouds are different as when compared to on premises. So one of the thing they they they they don't care is, uh, most of the Most of the institution don't care is that, uh, you know, they have to optimize their applications. They have to make it cloud ready kind of application so that they can use them best out of, uh, those resources while keeping the costs low. So I will be more looking into optimizing the application first itself and then looking into the, uh, cost effective solution because in cloud, they are alternate for everything. You know, depending on your, uh, use case, there are alternates, uh, which can suit your budget. So I'll be looking onto those alternate resources. Either it could be AWS, Azure, or anything. It's not a big thing. Fine. So it I'll just I'll just look out for the relevant cost effective solutions for achieving the same goal. So the my 2, Again, to revise my 2 primary, uh, uh, target areas will be, uh, getting that application optimized for cloud. Uh, no lift and shift. Just optimize the application for cloud. Spend some time brainstorming. And the second thing will be to, uh, basically, uh

Uh-huh. That that, uh, that question points to the basically, that question points to the to to the importance of having, uh, some sort of vault for managing, uh, sensitive data. If it was Azure, I would have used Key Vault. Uh, for the certificates, we'll be using things like and that sort of stuff. So, basically, uh, we'll be finding the relevant solutions, which can, uh, which can maintain the integrity of, uh, sensitive data. And then introduce, uh, those into the respective environments through the mechanism. Like, for example, EPlay calls or the integration from the platform itself. So so again, I'm using vaults and, uh, uh, basically, you know, the the the solutions which are therefore, keeping the sensitive data and integrate them with the with the with the platform so that, you know, there's no, uh, there's no, the nobody can look at those and they are not exposed to the external involvement or anybody who's even, uh, managing those stuff. So yeah. That that would be our goal

I don't have much Python experience. But, uh, from understanding this code, I think it will always cause an error. The The reason is that we are passing dev in the environment, and the condition says that if environment not in staging and production. So if you have given anything except staging and production, then it will always There is some sort of it will basically throw an error kind of thing. So it will always throw an error. So

I think, uh, the networking you are talking about here is in terms of the networks, not the social networking. Because I think manager has, uh, you know, if if it's about manager, then then you know, just technically His context is somehow also some social aspect social aspect involved into the job as well. But I I'm taking it from the from the digital, uh, computer networking side. So for a DevOps manager, he needs I think it's a it's, It it he should have, uh, networking experience. And, uh, one of the reason why, uh, I say is that, Uh, most of the most of the most of the issues we deal at the moment are somehow related to networking. Cloud itself, Based on the backbone of the cloud, it's all networking. Everything is in or on a remote location and network without network, it's impossible to imagine about cloud. And the issues we might face are very, very complicated you know, based on based on kind of, uh, the situation. It could it could be a simple networking firewall, uh, related issue, or it could be, uh, things like, you know, CDN issues. I I remember once we were, working with an application and try to fix it. And it turns out that it is somehow related to, uh, what's called, I forgot its name, but these are there are some things which are actually which are kind of, Which are they're back one of networking all the networking, uh, across the Internet. Autonomous systems. Yeah. So that was related to autonomous systems, and We were lucky that we had a we had a very good networking engineer with us who had experience in these autonomous systems. Uh, so so basically, you know, if a if a developer if the develop DevOps manager is having good experience onto the networking side, then he will be able to understand and propagate this information to the right people because business people don't know much about the all these technical things. And one of the, uh, responsibilities of a DevOps manager is to basically, you know, Uh, I have both sides. He has both 2 phases. So 1 is for the technical side and one is for the business side. So he can be a very good, uh, bridge between, You know, these, uh, when it comes to explaining a situation so he can he can work as a very good, uh, uh, you know, translator or kind of a Uh, person who can actually explain business in exactly what terms the problem is without going too technical. And, uh, so so I think it is very important for a manager to have networking cloud and, uh, a kind of an application experience kind of

This is, uh, I don't know how to answer this question. The reason is that, you know, it is It's it's when I see DevOps, it's it's I see just I understand only 3, uh, 2 or 3 things. The first is the networking. You know, when it When it's DevOps, that means basically things related to the cloud and, uh, you know, how they are deployed and, uh, you know, they They are being released, that sort of stuff. So so networking is the first thing that should be, uh, you know, that it that comes first when it comes to these kind of, uh, roles. And and, uh, so, you know, the term the situations we deal with, the problems we deal with, Almost 50% are related to networking. There are some final rules needs to be opened. There's some, You know, the packet flow we have to observe and troubleshoot and applications being slow. So and also, you know, uh, even even Even even for creating a new solution on the based on the cloud, we have to consider networking because, you know, we have to, uh, we have to figure out its, The right resources with right right right type of stuff right type of stuff to use to, uh, you know, expose your application to the to the to the to the, uh, to the, uh, to the right audience. So networking is actually the key and, uh, definitely, it is one of the very he should anybody who's a DevOps guy should have Good experience on networking. I think DevOps is not just, uh, something that that can be it's not a role that can a fresher fresher can do that. I think the first requirement is that you should have Some experience into application development, system administration, networking, that sort of things, and also, you know, how From a solution design perspective also, how things are designed, how the solution application are designed. So when that is that that base or that background is, uh, there with someone, then he can he can take up, uh, the role of a DevOps engineer. So definitely, I think, Yeah. These are all the specs that needs to be covered from a skills perspective from, uh, by a by a