By Mehakdeep Singh
At its peak, in 2015, GO-JEK had a few hundred drivers. It was around this time, we were onboard as consultants for a relatively unknown, unheard of startup. We took up the challenge because the market potential was unlike anything we had seen. But that’s a story in itself. 3 years ago when we began consulting for GO-JEK, this was the very first command we ran; we all remember it vividly:👇
The legacy backend codebase was named Stan Marsh. And… it really was something. It was a monolith on which, over 18 months, 18+ products were going to depend.
Looking at the code, one could understand the evolution GO-JEK has gone through — the decisions taken then and the footprints of each developer. This was at a time when we weren’t experiencing millions of orders a day. It was the backbone. It was all we had.
Stan Marsh has seen and helped GO-JEK grow from a hundred odd drivers to more than 1 million registered drivers today.
Over the three years since that day, we extracted more than 100 microservices from this monolithic MVP. Step by step. Deploying it is still an occasion, and only a few brave folks dare to take that leap. 😂
But the time has come to put it to bed and let the new generation take over. I’d hate to see Stan Marsh go, but as with everything technology, change is apparently the only constant. Or so we’re told.
At this moment, I can’t help but look back at those days when Stan Marsh was a nuisance, an irritant we couldn’t get rid of. But today, as I write this, I remember everything it stood for — an untold story of our own evolution from couple of orders to a 100+ million orders every month.
So long Stan Marsh, and thanks for all the fish. ❤️
P.S: More to come on a detailed journey and steps taken to transition from Stan Marsh to microservices.
We have hard technical problems at GO-JEK and want as much help as we can get from engineers. We’re changing the digital economy of Southeast Asia and want the sharpest minds to work with us. Please check out gojek.jobs for openings.
