Rajith Attapattu

Today I spent the whole day at a large fintech with their platform team & Sr Dev setting up Open telemetry alongside the Randoli observability platform. We managed to provide a comparable APM experience to their existing Datadog usage (at 60% cost reduction as per quote) without any sampling (they were sampling before to save cost) The SLO monitoring & real time log monitoring was icing on the cake. The Sr Dev was able to set it up with minimal guidance. In two weeks we turn on our SRE Agent. Can't wait to see the results. #Observability #DevEx #OpenTelemetry #Randoli

@JayDelhiBoy @hathyogi31 It's not for the trolls to decide. Most of them have nothing going on for them. Arjun is being picked by sides and he's taking wickets and scoring runs at the top level. He might be average but not all 11 in a cricket team can be superstars.

@hathyogi31 People troll him because he is in the wrong field.

Sara Tendulkar and Arjun Tendulkar perfectly highlight how hypocritical society can be. Just open Arjun Tendulkar’s comment section, it’s flooded with trolls. People say things like, “You’re only fit to bring drinks on the ground,” and constantly put him down. Now compare that with Sara Tendulkar’s comment section. And let me be clear, I’m not saying she deserves hate. She’s a biomedical science graduate. But that’s not what she’s famous for. Her 9.5 million followers aren’t following her for nutritional advice or academic expertise. On the other hand, Arjun may not be at his father’s level as a cricketer, but at least he’s trying. He’s putting in the effort. He doesn’t even need to work hard financially; he could easily live off his father’s wealth, show off a lavish lifestyle, drive expensive cars, go on international trips every week, post luxury photos, and people would go crazy: “Sir, sir, one autograph please!” “Sir, you look amazing!” His Instagram following would skyrocket. But instead, he’s choosing a tougher path. The real point of this post is simple: social media runs on illusion and display. The more you pretend your life is perfect, luxurious, and constantly exciting, the more successful you appear online. Being real doesn’t really work here.

Happy for Deontay. But he doesn't seem like anything close to what he was when he fought Tyson Fury.

After a long time I'm watching an IPL game bcos of Vaibhav Sooriyavanshi. The last time I watched a game was when Sanath Jayasuriya tuned up for Mumbai Indians #ipl #cricket

I can attest to that.

@johncrickett Software engineering is a lot more than coding. The AI agents are solving some aspects of coding while creating a whole can of worms in other areas. That being said the productivity gains it brings if harnessed properly can definitely help you build faster.

Software engineers six months ago: "An engineering manager has to be technical. How would they understand what the team is doing if they can’t build software?" Software engineers today: "Coding is basically solved. Non-technical people can just manage AI agents and build software now." Which is it?

I don't follow IPL, but I checked the scores just to see how Vaibhav Sooryavanshi fared. He smashed a 15 ball half century. No surprises there. A once in a generation talent 🔥 #cricket #IPL

The Eurostar is about to leave the station. I love everything about trains. #trains #Amsterdam

@TheTechMaharaj I'm so glad we did.

@rajith_a Let's catch up today - I'll be in the solutions showcase post Keynote.

I made it early enough to sit a bit at the front. But every @CloudNativeFdn KubeCon I'm simply blown away by the sheer scale of the main stage. #KubeCon #Kubernetes #Amsterdam

@SumitM_X I fail to see how this question is relevant for what most developers have to do in their day to day job. Just genuinely curious.

A question that is being asked a LOT in interview these days : Why Redis is single-threaded but still faster than most multi-threaded databases?

@TheTechMaharaj I'm so glad we did.

@ramiyengar it was kind of you to drop by. Meeting you at KubeCon is becoming a tradition.

Kubernetes cost management is top of mind for many who think about Day 2 operations. We had a full room for the OpenCost maintainer track talk with some really good questions during and after the talk. The community had a fairly positive reaction to 1. Promless mode 2. Kube model separating usage from pricing. I was proud to represent the OpenCost community and a big thank you to Alex Meijer from the community for helping me with the presentation as well as Danielle Zeis Tamao Nakahara for their support. #Kubernetes #FinOps #KubeCon

@SaiyamPathak Lol, the amount spent on tokens are no jokes these days.

@rajith_a 2k a month can now be spent on tokens now 🤣

I reviewed 100+ Kubernetes clusters this year. The #1 problem wasn't security. Not scaling. Not networking. It was cost. Teams burning $50K/month on GPU nodes for AI workloads that run 2 hours a day. AI isn't expensive. Bad infrastructure is. What's the worst cloud waste you've ever seen?

@chaz_6 @__karnati For 80% of the use cases it works if you are running kubernetes

@__karnati How does NATS compare to these?

Kafka, SQS, RabbitMQ are three message queues that most use. All different. Picking the wrong one will hurt you in production. Here's a simple breakdown on all three 👇👇 1️⃣RabbitMQ: classic message broker - Designed for task queues and job distribution. - Messages are consumed and deleted. - No replay is provided. Best for: background jobs, email queues, RPC patterns. Latency: very low. Scale: moderate. 2️⃣Amazon SQS: managed, serverless queue - No infra overhead to manage. Scales infinitely. Pay per message. - At-least-once delivery. Best for: AWS-native architectures, decoupling microservices. Latency: milliseconds. Simplest operational model. 3️⃣Apache Kafka - Messages are NOT deleted after consumption. They're retained by time/size. - Multiple consumer groups can read the same data independently. Best for: event streaming, audit logs, CDC, real-time analytics. Latency is very low. Scale is massive. Operational complexity: high. Do you need a replay? Neither RabbitMQ nor SQS are not option. Once consumed, it's gone. Kafka can replay from any offset, any time. - If you need to rebuild state from your event history, use Kafka. - If you just need to process a job, RabbitMQ or SQS. When teams pick Kafka wrong: → 5 services, 10 message types, lo.w volume → Now running a Kafka cluster 'just in case' → 3 engineers spend 20% of their time on Kafka ops SQS should do it in 10 lines of code. Pick boring where boring works. Start with SQS. Migrate to Kafka when you need replay or fan-out at scale. What's your team using and would you pick the same today? 👇

In my previous job spent 9 yrs building high volume message brokers and clients. In my current company we have Observability customers running all 3 in production. The reality is for 80% of the use cases I have seen any of those technologies will work. The decision should be based on availability of skill set, operational overhead and budget. Scale and volume is rarely a consideration.

@twtayaan Proper Observability is more than metrics, traces & logs. It's the ability to correlate across signals extracted from raw telemetry without having to stare at dashboards all day. Faster correlation is key for reducing MTTR.

🔍 What is Observability? Most people think it’s just monitoring. It’s not. Observability = knowing what’s happening inside your system without guessing. It answers 3 simple questions: 1. What is happening? 2. Why is it happening? 3. How is it happening? 1. Metrics: → What’s happening (CPU, memory, latency, traffic) 2. Logs → Why it happened (Errors, events, debug info) 3. Traces → How it happened (Request flow across services) Metrics = What Logs = Why Traces = How If you don’t have observability, you’re just guessing in production.

Security threats can come from anywhere including the very tools we use for identifying vulnerabilities. Recently trivy github actions were compromised. I applaud @AquaSecTeam for keeping an open source version of the trivy scanner available and hopefully won't get discouraged by this incident. For more details see the 1st comment.

Walked around Amsterdam last night #KubeCon