Cloud Migration
The 4-Week AWS Migration That Paid for Itself
Most migrations don't fail on the technology. They fail on scope creep, half-decided cutovers, and the quiet assumption that the plan will survive contact with production.
This is the story of one that worked. Four weeks, two applications, a set of databases, and every byte of user data and file storage moved off a struggling VPS and onto AWS for Siyakha. By the end, the platform was handling eight times more concurrent users than before, and the monthly hosting bill was 40% lower than the VPS one.
Our approach
The educational platform, one of Siyakha's projects, lived on a VPS. For a small user base, that setup is hard to argue with.
Then the user base grew which led to more activity and more load that the infrastructure could not handle. The traditional fix — a bigger VPS — meant paying for hardware that sat idle between the peaks.
The platform worked. It just required more resources to scale which meant a bigger bill.
Before configuring a single AWS resource, the team secured AWS credits. These credits fully covered the migration process and several months of subsequent hosting for Siyakha. This financial buffer eliminated a common migration pain point: no double hosting costs.
The challenge
Four weeks is the outcome of scoping the work honestly, and it is only realistic when three things are true: The applications can be containerised without a rewrite. If week one reveals the app needs restructuring first, the timeline changes. The data can tolerate a replication-based migration: we set up the new databases, replicate continuously, and cut over once the data has caught up. Someone on the client side can make decisions quickly. Most multi-month migrations aren't always slow because of engineering. They're sometimes slow because decisions need to go through several review processes before approval. If those three hold, four weeks is genuinely enough. If they don't, better to find out in week one than in week three.
The solution
Pyxis partnered with Siyakha to architect and execute a seamless migration of yowza!© to a secure, enterprise-grade AWS environment. We replaced the platform's failing legacy AI with a fully responsive bot powered by Amazon Bedrock and implemented CloudWatch and Grafana for proactive, real-time issue detection. By integrating the PayFast payment gateway, automating the testing framework to relieve the stretched internal team, and establishing robust deployment pipelines, we transformed yowza!© into a resilient, scalable platform ready to support tens of thousands of active SMME users.
Measurable results
8X
Concurrent Users
4 Weeks
Kickoff to cutover
-40% /pm
vs the previous bill
The Phase 1 migration of the yowza!© platform has been a highly successful and necessary step for Siyakha. Godfrey and the Pyxis team didn’t just move us to AWS; they fundamentally stabilised our infrastructure, implemented robust security controls, and finally gave us the live monitoring and alerting we previously lacked.
Loni Hutchison,
Head of Digital
