hyperswitch

fcd94619 - fix(ci): resolve Xvfb display conflicts in parallel Cypress execution

fix(ci): resolve Xvfb display conflicts in parallel Cypress execution ## Problem When running 6 parallel Cypress test batches across 2 connectors (12 total processes), each Cypress instance was attempting to spawn its own Xvfb server, causing display conflicts: - "Display :105 is already in use and the reuse option is false" - "xcb_connect() failed, error 1" - GLDisplayEGL::Initialize failed ## Root Cause 1. Each of 6 parallel Cypress instances tried to start separate Xvfb servers 2. Xvfb requires ~2 seconds to initialize after starting 3. Race condition: Cypress instances starting before Xvfb was ready ## Solution 1. Start ONE shared Xvfb server on display :99 before tests begin 2. Add wait loop (up to 15 seconds) using xdpyinfo to verify Xvfb accessibility 3. Set DISPLAY=:99 in both shell environment AND step env block 4. Applied to all 3 jobs: runner, runner_alpha, runner_v2 ## Verification - Tested in Docker (ubuntu:24.04) which matches GitHub Actions environment - Confirmed Xvfb becomes ready in ~2 seconds - Verified 6 parallel processes successfully connect to shared display - No display conflicts observed 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>