Splunk Platform Modernization

Client Context

Splunk's core search dashboards ran on legacy Backbone.js and jQuery, with fragmented visualization libraries, code bloat, and regression risk across on-prem Splunk Enterprise and Splunk Cloud, plus an enterprise accessibility audit deadline for 120+ dashboards. The work ran directly with Splunk's product engineers, PM, and UX.

Execution

Used a Strangler Fig pattern with mounting bridge adapters so Backbone views and React components could coexist, and built AST codemod tooling (Babel, jscodeshift) to convert views automatically. Added route-level lazy loading, code splitting, and cache-aware data fetching, upgraded Highcharts from v4 to v6 with accessibility modules, and put contrast and keyboard behavior into design tokens.

Outcome

Hundreds of legacy pages converted with under 5 production regressions, and the playbook was adopted by 30+ engineers. Time-to-interactive improved 30% (2,800 → 1,900 ms) with parity across Splunk Enterprise and Splunk Cloud, and 120+ dashboards reached full WCAG 2.1 AA compliance ahead of the audit deadline.

Why It Mattered

Large migrations succeed when the right path is the easiest one: codemods did the mechanical work so engineers could spend their judgment where it counted. Building accessibility into design tokens, not last-minute patches, kept the dashboards usable for people who navigate by keyboard and screen reader long after the audit.