Pranay Prateek

ohyaml.wtf | YAML Quiz YAML is known to be nobody's friend and almost everyone's enemy. Try this to see if it's your friend or foe!

give it a shot - ohyaml.wtf

devops/sre friends: how fluent are you in YAML—on a scale from “2 spaces” to “tabs”? 😈

we handcrafted a cursed little trivia to make you go what-the-colon. any arcane YAML we missed?

think you cracked it? flex (or confess) your score in the comments 🤣

have a fun friday 🥳

Why Observability Isn’t Just for SREs (and How Devs Can Get Started) This blog is an attempt for anyone lost to find their way into observability and a wake-up call for devs to they should think about observability more actively today than ever before.

Why Observability Isn’t Just for SREs and How Devs Can Get Started - signoz.io/blog/why-obs...

For example, here's a dashboard from SigNoz on how you can monitor a RAG system based on LLama Index. Isn't understanding how systems you have built equally important as building that system?

We recently wrote a guide on how devs can get started with observability (attached in comments)

the real value of an engineer is taking a code and actually deploying it to solve problems for customers.

And for that understanding things like deployment, observability, etc. are super critical

It's just that in colleges, people are taught that writing code is what developers do.

Code is easy—thanks, AI.
Keeping it running is the real superpower.

It's surprising for me how many junior developers never hear about observability..

In this LLM era, when writing code has been made much easier by tools like Claude,

Trace Funnels Trace Funnels provides visibility into the step-by-step progression of requests (spans) through your distributed system. By defining up to three sequential spans within the same trace, you can…

More details about Trace Funnels - signoz.io/docs/trace-f...

including funnel validation, step-by-step analytics

Always grateful to the dev community for their continued support 🕺

Let’s keep pushing the boundaries of what’s possible in open-source observability. 🚀

Crossed 23,000 Github stars over the weekend 🥳 🎉

We are continuing to define the standards in open source observability.

One feature I am particularly excited about is Trace funnels where you can analyse user journeys and conversion funnels directly from your traces,

Integrating Vercel AI SDK Telemetry with SigNoz Enable telemetry in your Vercel AI application with OpenTelemetry and send traces, logs, and metrics to SigNoz

Docs - signoz.io/docs/llm/ver...
Sample Repo - github.com/SigNoz/verce...

What AI tools in your stack need better observability? Drop them below and we'll explore solutions.

🔧 Why this matters:

- Unified telemetry across AI stack
- Better debugging for agent handoffs using a open standard
- Standard traces, metrics, and logs

📊 The setup is surprisingly simple - just a few lines of code to get full observability into your AI workflows.

OpenTelemetry is becoming the lingua‑franca for LLM app telemetry. Multi-agent systems need standardized observability.

Here's how to monitor your Vercel AI SDK app with OpenTelemetry + SigNoz in minutes.

it certainly will survive 2030 :). - their AI models are one of the leading ones

From Sequential Bottlenecks to Concurrent Performance: Optimizing Log Processing at Scale We optimized log processing pipeline by moving from sequential to concurrent processing at the entry level, achieving 30% higher throughput and better resource utilization without increasing…

details: signoz.io/blog/optimiz...

Before :
CPU Usage: 2,694m/4,000m (67% utilization)
Processing Rate: 35.7k logs/second

After:
CPU Usage: 4,352m/4,000m (108% utilization)
Processing Rate: 46k logs/second (with 66k peaks)

🔥 SigNoz team is on fire! We shipped a core improvement recently, which led 30% higher log processing throughput without adding any infrastructure.

We moved from sequential to concurrent processing architecture and that have us some good perf improvement

Platform Engineer Looking for a Platform engineer to join our team at SigNoz with expertise in handling Kubernetes at scale and expertise in Go

JD and application link - jobs.ashbyhq.com/SigNoz/01ebd...

Also, we are fully remote.

PS: If you know any one in your network or friends, please send them my way. Good karma & all.

🚀 We are looking to hire platform engineers in the US at SigNoz. If you have managed huge k8s clusters and like handling scale, do reach out

22k+ ⭐ in github repo. funded by YC and top VCs in Bay Area - scaling fast & looking for folks who have experience handling scale and have experience in Go

Great to see increasing number of developers finding SigNoz's single pane observability valuable 🚀

If you are still running prom-grafana-loki stack, may be good time to check SigNoz out :)

You can either self-host it or just use SigNoz cloud

YC-backed Shaped AI Swapped a Siloed Toolset for SigNoz, achieving One-Stop Observability YC-backed Shaped AI Swapped a Siloed Toolset for SigNoz, achieving One-Stop Observability

full case study: signoz.io/case-study/s...

After: "Every single time we have an issue, SigNoz is always the first place to check"

- Single dashboard for logs, metrics, and traces
- Predictable, cost-effective pricing

The migration? Just three lines of configuration change, thanks to OpenTelemetry!

Before: Fragmented stack with CloudWatch + Honeycomb

- Constant context switching between platforms
- Unpredictable pricing causing developer anxiety
- Slow troubleshooting workflows

Great to see how AI-native companies like
Shaped are finding lot more value by moving from multiple, siloed tools to a single pane for observability with SigNoz 🚀

MCP Observability with OpenTelemetry Learn how to monitor MCP server-client systems with OpenTelemetry. This guide covers the importance of observability, common failure modes, and a step-by-step implementation to ensure your MCP server-...

signoz.io/blog/mcp-obs...

Distributed tracing with OTel is the perfect tool to solve this problem.

1/ Context propagation across service boundaries
2/ Vendor-neutral, open standards
3/ End-to-end traceability from agent prompt to final output

here's a detailed guide. Let us know what do you think?

You prompt an agent, a tool gets invoked, response generated.

But what happens in between?

And when something breaks, how do you debug it?

Agentic AI is the rage these days, yet most teams are flying blind when it comes to observability in MCP architectures.

They need to know if you have a deeply-thought-out, credible plan to win. They're testing your conviction under pressure.

We gave our answer. The call ended. And then... silence.

A few days later, I woke up to a missed call from a US number at 6 AM.

"DataDog is a massive public company. How are you possibly going to beat them?"

This is a critical lesson for any founder.
YC already knows your market; they don't need you to explain it.

Because a crisp, clear promise is more valuable than a dozen half-baked features.