Lead Technical Account Manager @ Dynatrace · Bengaluru, India
Observability engineer. Tool builder. Teacher. Still a student.
LinkedIn / The Harithsa Notes / npm / DEV
Hey, I'm Vishruth. I work where production systems, customer problems, and curious people meet.
Sometimes that means tracing a failing transaction. Sometimes it means giving an AI agent access to the right evidence. Sometimes it means turning a cloud workshop into a chai startup with outages, bill shocks, and a checkout murder mystery.
The thread connecting it all: understand the problem, build something useful, and leave enough context for the next person.
~$50M in business our team helps protect at Dynatrace.
My role is to help customers turn observations about their technology into business outcomes. That means connecting application health with user journeys, analyzing usage patterns, and helping teams understand where reliability and experience affect the business.
The questions I care about: Who is affected? What are users trying to do? What does the business need us to improve?
My personal projects range from AI instrumentation to money management, books, and small-business tools. A few doors into the workshop:
| Project | What I wanted to make |
|---|---|
| FinBuddy | A personal finance app with budgeting and an AI assistant informed by Arthashastra principles. |
| Manahsmriti | A home for personal projects, starting with a searchable book catalog and private reading libraries. 223 catalog entries documented so far. |
| SalesDrone | A sales workspace that keeps enquiries, conversations, and the next follow-up together. |
AI agents need observability, too. I build tools that make their calls, dependencies, and failures easier to inspect.
| Project | The question behind the code |
|---|---|
| MCP × OpenTelemetry | What actually happened inside that tool call? Automatic MCP instrumentation, trace context, and OTLP export. |
| Copilot SDK × OTel | Where did the agent spend its time and tokens? Model calls, tool spans, token metrics, and latency. |
| Dynatrace Managed MCP | How can an AI assistant work with the Managed environment a team already runs? 39 documented tools, with telemetry built in. |
| Local LLM × OpenLLMetry | Can a self-hosted chatbot be observable end to end? Open WebUI, Ollama, and a traced backend exporting to Dynatrace. |
| dt-mac-agent | What is my Mac doing while all of this runs? A lightweight macOS metrics agent and a Dynatrace health dashboard. |
ChaiCart Cloud Workshop turns cloud concepts into a business people can build, operate, and occasionally break.
2 days · 24 teams · 1 chai startup
Workshop design: 120 learners in teams of five.
Build it on day one. Keep it alive on day two. Work through outages, cost surprises, and “Who Killed Checkout?” using logs, metrics, traces, and evidence.
The learning continues in InspiredLearning, a companion portal for workshop participation, results, and certificates. I also contribute and learn through OffOn and share practical Dynatrace sessions with customers and partners.
I write about how systems work, what changes when AI joins the workflow, and what a career in technology teaches you about yourself.
- Agentic AI for legacy — connecting AI assistants to existing Managed environments.
- The War Room Is Dead — the move toward AI-assisted incident investigation.
- Chasing Greatness — Ayrton Senna, craft, and finding your own direction.
The human behind the telemetry
My path into technology wasn't a straight line. I dropped out of engineering college. I kept building, asking questions, and learning from people willing to share what they knew.
That shapes how I teach: give people something real to try, room to get it wrong, and enough understanding to try again.
Away from the terminal: gaming, reading, and thinking about how to live with a little more purpose. Some of that finds its way into my writing. Some of it becomes another side project.
I'm interested in observable AI agents, practical OpenTelemetry tooling, and learning experiences that get people building.
Have a repo, a strange trace, or an idea for a hands-on workshop? Start a conversation.
Build. Instrument. Explain. Repeat.