Infrastructure

JailCore's existing infrastructure incorporated Kubernetes clusters and a Postgres database. However, the setup relied on outdated versions of these components that lacked the optimizations available in more recent releases.

Database and Rails n+1

DBA pinpointed an "n+1 problem" as a major performance hindrance in the JailCore application. This is a common challenge in Ruby on Rails apps. It arises when an app conducts excessive database queries, inflating the load on the database and computational resources. Additionally, the JailCore app used an older version of Postgres, which lacked the performance-boosting features of newer versions.

User Dashboard

One of JailCore’s ambitious offerings was a customizable dashboard, which although innovative in conception, came with its own set of challenges. Previous development left the dashboard with an unfinished feature that, in theory, allowed users to modify the layout by ordering, positioning, and resizing elements. However, resizing didn’t work, the widgets only displayed dummy data, and users could not save their personalized dashboard states.

The Solution

The DBA team worked closely with JailCore's CTO, Robby Clements, to ensure a seamless exchange of insights and product context. This collaborative approach not only accelerated development but also ensured that solutions were precisely tailored to JailCore's specific needs. The real-time collaboration between the teams facilitated quick decision-making and the rapid implementation of critical improvements.

Optimized Data Operations

The DBA team leveraged their Rails expertise to develop a comprehensive solution to the n+1 problem. By shifting some of the processing functionality from the application layer to the database layer and utilizing database views, they significantly reduced the time required for query execution. Additionally, batch insert processes were introduced, replacing time-consuming one-by-one data imports. Caching mechanisms were also implemented, minimizing the load on the database. Moreover, the upgrade to a newer Postgres version introduced significant optimizations to index management and JSON queries.

These many improvements mitigated the Rails n+1 problem, reducing excessive database queries and enhancing the platform's overall responsiveness.

Modernized Infrastructure

The DBA team upgraded JailCore’s Kubernetes and Postgres versions. This essential overhaul was completed within a tight timeframe, and ensured that JailCore's platform was running on modern, optimized infrastructure. In particular, the migration of Postgres to a newer version not only addressed existing inefficiencies but also primed the system for future scalability.

Robust Dashboard

DBA was able to identify and resolve bugs that prevented some important functions from executing. These fixes, along with earlier data optimizations, cleared the way to complete old dashboard features, as well as develop new ones. Through real-time data access (web sockets) via dashboard widgets, users gained instant visibility of facility operations. They could also save customized widget states, allowing them to tailor the dashboard to their individual roles.

Seamless Integration of Diverse Technologies

A significant challenge was ensuring the smooth integration between the back-end, driven by Rails, and the front-end. To facilitate the new configurable dashboard, React - a front-end JavaScript library used for part of the front-end - needed to be updated to a newer version. However, the update broke the React integration with Rails. To resolve this, the team updated an app dependency that facilitated communication between Rails and React.

With the integrations working properly, the team implemented two approaches to presenting data on the front-end, both leveraging the power of the Rails back-end framework.

1. ERB (Embedded Ruby): This method embeds Ruby within HTML. It's used for more static parts of the application, where content remains largely consistent. With this approach, the back-end sends HTML embedded with Ruby code in response to user interactions.
2. React with JSON: For sections of the application requiring greater flexibility and user interaction, the back-end sends raw data in a JSON format using websockets and HTTP calls. This approach enables the React-powered front-end to determine the optimal presentation of the data to the user.

By integrating these techniques, DBA ensured that users enjoy a dynamic and seamless experience, regardless of the underlying technology.

Featured Tech for this Project:

Ruby

Rails

PostgreSQL

React

JavaScript

Google Cloud

Redis