The Night My Store Crashed During Diwali Sale — and What I Learned

It was 11:04 PM on Diwali Eve when Rohan, our backend engineer, called me. I was sitting on the floor of my living room, watching the orders pour in on a laptop, eating mithai my mother had made. We had crossed ₹18 lakhs in the first forty minutes. I had never felt more proud of anything I had built in my life.

"Bhai," Rohan said. "The site is down."

I remember staring at my laptop screen. The order counter had stopped. The little green indicator I had been watching like a heartbeat — gone.

What Downtime Actually Feels Like

People talk about downtime in percentages. Three nines. Four nines. SLAs and incident reports. That is not what downtime feels like at 11 PM on Diwali when you have spent your family's savings and three years of your life building something.

Downtime feels like standing outside your own store and watching the shutters come down while a queue of customers stretches around the block. It feels personal in a way that is hard to explain unless you have lived it.

We had approximately 12,000 customers on the site at the moment we crashed. Our hosting provider's shared infrastructure had buckled under the flash sale load. We had not bought enough capacity. We had not tested at scale. We had assumed things would just work.

They did not.

The Reckoning

By 3 AM we were back online. By then we had lost ₹40 lakhs in abandoned orders — we could see every single one of them in our logs, like gravestones. Customers who had tried to complete checkout six, seven, eight times before giving up. One customer had tried fourteen times over ninety minutes.

Fourteen times. She wanted to buy from us fourteen times and we failed her fourteen times.

That number haunted me for months.

The next morning I sat with Rohan and our product lead and we made a list of everything we had compromised on to save money or ship faster. The list was long. Shared hosting instead of edge infrastructure. A checkout flow with no fallback. Session state stored in a way that did not survive restarts. No queue for order processing. We had built a beautiful storefront on a foundation that was never meant to carry real load.

The Decision That Changed Everything

I will not pretend the rebuild was easy. It took four months and it cost us more than we had planned. But the conversation that started it was simple: I asked Rohan what it would take to make sure we never had a Diwali night again. Not just handle traffic spikes — actually never have a night like that again, ever, regardless of what happened.

His answer was edge rendering. Serve the storefront from everywhere, not from one place. Make the site stateless and fast by default. Let the infrastructure absorb the spike instead of buckling under it.

When we eventually moved our storefront to QuantumOS X3, the thing that struck me was not the admin panel or the features. It was the first time I looked at the latency dashboard during a sale and saw a p95 TTFB under 100ms. From Chennai. From Delhi. From Tier-2 cities where our customers are. The store was fast everywhere, not just in a data center somewhere.

The Next Diwali

Last Diwali we ran a sale. Bigger than the one that had crashed us. More SKUs, more discounts, better marketing. We crossed ₹18 lakhs in the first twelve minutes.

By midnight we were at ₹1.2 crore.

The site did not crash. The dashboard did not go dark. Rohan sent me a message at 11:04 PM exactly — he had set a reminder — that just said: "We're good."

I cried a little. I am not ashamed to say that.

What I Would Tell Every Founder

The lesson is not "use better hosting." The lesson is that infrastructure is not a cost center — it is the promise you make to your customers every time they try to buy from you. When it breaks, you are not losing revenue. You are breaking a promise to someone who chose to trust you.

Build infrastructure worthy of the trust your customers are placing in you. Everything else is commentary.

Subscribe to the QuantumOS Dispatch — weekly insights for commerce operators who want to compound their advantages.