Shantanu Bose

Could you help me understand more about your background? But before you so I have a decade old experience, uh, in IT industry and out of which a relevant experience of 6 to 7 years in DevOps, worked on different clouds like AWS, Azure, with different technologies like Kubernetes, Docker, basics of AWS, EC 2, S3, and a lot of that lot of the deployments, production environment, testing environment, Still a lot of CICD things and everything. So most of the things I have almost touched, uh, in DevOps, this whole vast career of mine to DevOps. So, yeah, this is it. Uh, starting with entrepreneurship, ended now into DevOps. So it's been a whole around 11 years journey.

And handling last databases. I'm not pretty much sure with the Python thing. Uh, I know basics of Python, but I'm not pretty much sure about it. So we still try to take a crack into it. I'm not considering Python to it, but, usually, when we work with the large databases, we try to do some uh, distributed computing on them. So that's how we try to process big datasets with which we usually do with Hadoop or Spark. I'm not pretty sure with Python. Maybe we'll be using Pyspark for that.

I recently did that, actually. I mean, not recently, but, yeah, for a long period of last, maybe 8, 9 months, I'm doing that. So we are using Prometheus for it, uh, Prometheus Grafana to do the monitoring of our ports in Kubernetes. Uh, so, yeah, we tried doing Fluentbit also. So we get the logs while using Fluent Bit, and we finish real time methods. We get it from, uh, Prometheus and showcase them on Grafana.

What's an address? So if it's a multi multi cloud deployment, I would definitely suggest go for Terraform. And if it's in premise or not exactly in premise would be the right word, but it's something which is very much of AWS thing, then we can definitely go for AWS CloudFormation. But, yeah, definitely, Terraform has its own, uh, value add to everything actually, taking from state file to plan, apply a lot of stuff into Terraform, where CloudFormation gives a very much beautiful view of a designer thing where we are writing a simple YAML or JSON Terra CloudFormation, and we can see that, okay, uh, how this whole architect is going to look like. So that's that's upper end in CloudFormation. But, yeah, obviously, if it's we have a multi cloud deployment, then definite CloudForm is the perfect fit over there.

Well, uh, you need to automate a regular task that must be performed in the Linux servers with scripting language. Yeah. If it's a regular task which needs to be performed on a Linux server, uh, I I would choose Python, uh, maybe at for a very certain interval of time that you're okay with. Maybe maybe after every day at 12 PM, it will run. And if there is architecture behind it, maybe it's something, so I would go for Lambda functions in Python that can do that. Uh, but but, yeah, in this scenario, CronJob is the best thing to do. Batch scripts can help

What steps do you take from the root cause analysis when your application behaves unexpectedly and how to prevent it from happening in the future. See, uh, from my scenarios, when I work on a product, actually so the first thing first is that every time something is happening not expected it has to be, we try to check the root cause by finding by going through the architecture. K. Okay. Fine. Let's suppose I have a port and I have a product, and I know that, Keith, this product is going to run into a namespace. Right? And that namespace gets pinned up every time, uh, by by by, uh, API. Right? And that API is made up of some help charts behind that. And so I tried to look into the YAML file first. Okay. The YAML file, how this YAML file is going to make this bot look like if it's all good. If it's not all good, what are the configurations? Why maybe it might be sometimes possible that it's consuming more memory, more CPU. Sometimes it's like, okay. Jobs are not starting because the port configuration is very weak. It's only saying that, okay. You can have 1 GB of memory, and the port configurations to start, it needs 1.5 GB of memory. So these are the small things. So but I personally feel every time, if you have a clear architecture ready with us, it's very easy to look into where exactly things are going

Consider the following. We need to continue to confirm the performance part part that could prove successful implementation. Uh, replica 3 selector will check. Match labels, Not sure, but maybe the match labels, uh, 2 different labels, uh, in the next one. Apart from that, uh, without without maybe maybe if indentation is correct, I feel the app labels are wrong. Uh, that could be one of the reasons. Rest, not sure. Rest I think this is the only thing.

Given the below Python code, explain what the issue might be. What the issue might be? Uh, code looks good to me. Try catch. Try accept. The final deploy for cleanup. Not sure. Uh, the code looks good to me. I have to I'm not pretty much sure that there's a function which is doing this. Maybe the main file, it's it is not returning any value.

How do I do version control and deployment in multiple environments to ensure consistency across production? Yeah. So it's version control, we basically use Git for that, and all our deployments have the Git Git versions with them, actually. So that's how we are controlling our version control for different deployments, and we have our QA and prod deployments at different production numbers. And, accordingly, we do that. We uncheck on a very diff various multi cloud deployments also. Right. We use Git for that, actually. But it's, like, 1, uh, 1 minute 20 seconds for this question. I don't know. How you handle version control and deployment in a multi cloud environment to ensure consistency across production and production setups? Version control is Git only. Yeah. We'll be using CodeKit for that.

When configuring the infrastructure, how do you ensure that the system is securing cybersecurity threats? First thing first is our security groups. 2nd, then we have it's depending upon how the app or the product is exposed to a web, then that then comes to a lot of stuff, actually. If it's a simple database infrastructure, we can talk about NAT gateways, how the NAT gateway is coming, how ingress is getting con getting connected with the web, how where the e c two is maybe in so the public subnet. So we try to make sure creating a private subnet that web gets access to the private subnet via bastion host. A lot of stuff can be ensured to make sure that the data is secure against cybersecurity threats.