Check Out Sanil Khurana's LinkedIn Stats (Last 30 Days)

Hi folks, I recently began reading and understanding Authorization systems and it's amazing how much complexity goes into building a good Authz system. I distilled all my learnings in a 15-minute blog that covers RBAC, ReBAC and ABAC here - https://lnkd.in/gePX2M-m It's a fascinating topic with tons of interesting software engineering and I'd highly recommend checking it out.

Why doesn't NGINX work for Cloudflare in handling over 1 trillion requests a day? 🌐 Cloudflare recently shifted to using their custom-built tool Pingora from NGINX and wrote an amazing blog about their journey. Firstly, the number is MASSIVE, 1 TRILLION requests / day. That's absolutely crazy and probably the sort of problem only Cloudflare can have. But why didn't NGINX work? 1️⃣ Lack of Specific Functionalities: Cloudflare needed specific features that NGINX didn’t offer. For example, they wanted the ability to send a request to a different origin server with a different set of request headers when retrying or failing over a request. NGINX didn't have this functionality out of the box. 2️⃣ Performance and Cost Issues: The NGINX architecture made it less performant and more expensive. In NGINX, each worker is a separate Linux process, meaning each has its own data and connection pool. Scaling up workers requires adding more connections for Cloudflare and their origin servers, leading to increased costs and performance issues due to additional TLS handshaking between Cloudflare and the origin servers. 💸 3️⃣ Memory Safety Concerns: They had concerns about memory safety issues with C and wanted to move to a more secure language, like Rust. 🔒 The blog post is a great deep dive into the challenges Cloudflare faces and how they alleviated them with Pingora. 🌟 https://lnkd.in/gK-MuDVj Highly recommend giving it a read to understand more about Cloudflare’s impressive journey and innovations! 🚀 If you enjoyed this summary, follow me on LinkedIn for more insights and updates!

How does Booking.com do observability with tens of millions of custom Events per second? 🤯 I recently read this amazing article by Santanu Sahoo on how Booking.com built custom events to build an observability stack for logs, metrics and traces. 🌟 Here is a quick explanation on how they achieve this at such an impressive scale: 1️⃣ Events as Units of Work: A single event can represent an individual HTTP request, a single cron job, or similar tasks. Each event captures the complete context of the work done. 2️⃣ Detailed Information: Each event includes crucial data about the work, such as latency metrics, whether the request was successful, performance characteristics, runtime environment, and more. 3️⃣ Kafka for Event Processing: Events are sent to a Kafka cluster, where specialized consumers process different types of events and store them in appropriate datastores (like ElasticSearch for logs and Honeycomb for traces). 4️⃣ Actionable Insights: Engineers use these datastores to gain insights into user behavior, monitor the health of their services, and track key performance indicators (KPIs). 📊 The scale at which Booking.com operates is truly mind-blowing! Processing tens of millions of custom events per second to maintain observability is an incredible technical feat. 🚀 Read the full article here - https://lnkd.in/g_UFDpVu If you enjoyed this post, follow me on LinkedIn for regular insights and updates and let's upskill together! 🚀

This Github repository is the best way to learn software design patterns and system design concepts (It has more stars than MongoDB's github repo!) 🌟 It covers everything, from software design patterns for LLD for almost every language, microservices, and cloud architecture, serverless systems for system design, and even some resources on machine learning. One of the interesting posts I read from this repo this week was "Architecting for Reliability" - https://lnkd.in/gG3vbtkk. It's about what it means when we hear the word availability and reliability, how to achieve high levels of availability, and most importantly, what it costs. 💡 Check it out here - https://lnkd.in/gQF7vnmQ Follow me as I read engineering blogs and whitepapers to level up our skills together! 🚀

It's crazy how the App team at Quizizz managed to do the impossible - 1️⃣ They slashed the update size by 99%! 2️⃣ Resulting in download time plummeting from 1 minute to just 1 second! 3️⃣ Adoption time shrunk from 3 weeks to a mere 3 days! All of these are crazy reductions impacting tens of millions of users. But how does it work? The idea is relatively simple, instead of downloading the entire updated app, they used a couple of utility tools to just download the changes in the JS bundle code in a background process in the application. My friend Bhavya Rawal wrote a blog post detailing the entire process of making code changes, and how it is deployed to users. Its a great read to understand how teams can release updates quickly! 🚀 - https://lnkd.in/ghXPndbm Follow me for amazing resources to level up your engineering!

🔍 Indexing in databases is simple and highly effective to improve query performance, but are you sure you aren't overusing indexes? This eye-opening blog post by Jeff Kaufman was a short but amazing read on when people overuse indexes, and end up causing more harm than good - https://lnkd.in/gDUKa4k4 I love this short blurb that describes the intent of the post perfectly - 'This post is not anti-index, it is anti-"you should never be doing full table scans in production"' 💡 Want to dive deeper into discussions on optimizing database performance and more? Follow me on LinkedIn for constant resources as I share the most interesting tech articles I read every week and let's level up together!