Built in 2021
Daycohost Costumer Portal
Customer Portal
My tasks were as varied as creating the views, backend development, integration between web services and integration between enterprise systems.
Made with
Links
A hosting customer wants to pay an invoice, check a server and open a support ticket. At most hosting companies, those three tasks live in three different systems. Billing runs on one platform. Server management runs on another. Support tickets sit somewhere else.
Daycohost wanted one portal where customers could do all of it. No jumping between tools. No calling support for basic tasks.
I joined the team to build that portal.
One screen, many systems
The portal gives customers control over their hosting services. They can view and pay invoices. They can manage their server settings. They can open support tickets and follow them.
Behind the screen, the portal talks to several enterprise systems. It pulls billing data from one. It gets server status from another. It syncs support tickets with a third.
The customer sees one interface and never has to think about the systems underneath.
Work across the stack
I worked on every layer. The frontend uses React with Next.js for server side rendering. That means the server builds the page before sending it, so the first load is fast. Customers see their data right away.
The backend runs on NestJS, a Node.js framework built for TypeScript. With TypeScript on both sides, types were checked from the database to the browser. The compiler catches type mismatches, so fewer bugs get through.
I built a view for each main feature: invoice lists, server dashboards and ticket screens. Each view had to handle loading states, errors and live updates.
Adapters for every system
Integration was the hardest part. Daycohost already had its enterprise systems. The portal had to work with them, not replace them.
I built an adapter for each outside system. An adapter is a small layer that speaks one system’s language. It translates requests and responses into one common format. The rest of the app does not care where the data comes from.
Some systems used REST APIs. Others used SOAP, an older XML based protocol. One used a custom protocol. Each adapter hides those details from the main app.
Fast pages from slow systems
Enterprise systems do not always answer fast. A billing query can take several seconds. Customers cannot wait that long on every page.
So I added caching at more than one level. Recent data stays in memory. Older data lives in a shared cache. When data gets old, the system refreshes it in the background.
Customers see their information right away, and the data stays correct without the portal asking every system over and over.
When a system fails
Connect many systems and something will break. An outside service goes down. A response comes back malformed. A network call times out.
I handled errors across the whole integration layer, and each kind of failure gets its own response. Some errors trigger a retry. Others fall back to cached data. Serious failures alert the operations team.
Customers see a friendly message instead of a technical error. They know something is wrong without seeing a stack trace.
Lessons from integration
Connecting systems sounds simple. Real systems have quirks. Documentation is sometimes wrong. APIs behave differently from what their docs say.
I learned to write defensive code. Assume outside systems will fail, and plan for it.
I also learned what good boundaries are worth. When Daycohost updated one backend system, we changed only one adapter. The rest of the portal kept working.
Interested in working together? Check out my services and pricing to see how we can collaborate on your next project. Let's Build