Crisis rescue after lead dev departure mid-migration
A Dutch e-commerce platform with €2M ARR lost their lead developer mid-migration. Two junior devs remained, a critical infrastructure deadline was looming, and the system was unstable. We stabilised, completed the migration with zero downtime, and cut infrastructure costs 40%.
Challenge
What was broken
A Dutch e-commerce company generating €2M ARR was in the middle of migrating from a legacy PHP monolith to a modern React + NestJS stack when their lead developer resigned without notice. The two remaining developers were junior and had limited context on the migration plan. A critical go-live deadline with a major retail partner was 5 weeks away. The existing system was already under strain, with intermittent outages during peak traffic, and the half-completed migration had left the codebase in an inconsistent state.
Our Approach
How we thought about it
We sent two engineers on site within 48 hours — one backend, one fullstack. The first priority was stabilising the live system and stopping the bleeding. We identified three active bugs causing the intermittent outages and patched them within 24 hours. Then we spent 3 days documenting the existing migration plan, identifying what had been completed, what was broken, and what was missing entirely. We re-scoped the migration into phases, with the partner go-live deadline driving strict prioritisation.
Solution
What we built
We completed the NestJS API layer that had been left half-built, migrated the product catalogue and order management to PostgreSQL on GCP, and containerised the deployment with Docker and Cloud Run. The React frontend was completed and connected to the new API. The migration was done using a blue-green deployment strategy — the new system ran in parallel for 2 weeks before the final cutover, with real-time data sync between old and new databases. Zero downtime. We also wrote comprehensive onboarding documentation and ran two knowledge transfer sessions to bring the client's three newly hired developers up to speed.
Results
What shipped
The retail partner go-live happened on schedule with no incidents. Infrastructure costs dropped 40% compared to their previous hosting setup. The three new developers were independently productive within 2 weeks. The system has run without a critical incident for 4 months since the migration, and peak traffic handling has improved significantly with the new architecture.
System overview
“We were in full panic mode. Sigmalion showed up within 48 hours, calmed everyone down, and just started fixing things systematically. No drama, no blame, no excuses. They ran the migration flawlessly and left us with a codebase that our new team could actually understand.”
Got a similar challenge?
Let's talk about your situation — 30 minutes, no commitment, and you'll leave with a clearer picture of how to move forward.