Drupal Migration — Performance Monitor

Bottom line — Spectre_5 is running normally over the last 24 hours (sub-500ms response outside one brief overnight spike, apdex 0.74–0.88), and the week shows a traffic-volume anomaly on Oct 3 rather than a performance one. With Drupal 11 now carrying about nine weeks of production traffic since the Aug 1 upgrade, it is already outperforming every prior phase on both average response time and error rate, while apdex holds roughly flat with the Drupal 9/10 era.

• 24h: a brief overnight latency spike to 764ms at 02:09 CDT (low-traffic hour, apdex dipped to 0.74) was the only outlier; daytime response climbed to 633ms during the 09:00–10:00 CDT traffic ramp, in line with normal load behavior.
• 7d: request volume dropped sharply on Oct 3 (50.4K requests vs. 150–230K on surrounding days) while response time (480ms) and apdex (0.81) stayed normal — points to a traffic dip, not a backend issue.
• D11 early read: since Aug 1, Drupal 11 averages 591ms response and 0.21% errors — better than D10 (663ms, 0.36% err), D9 (699ms, 0.42% err), and the original D8 baseline (679ms, 0.39% err). Apdex is 0.77, essentially even with D9/D10 (0.77/0.78) and still a touch below D8 (0.79). The sample (~9.25M requests over ~9 weeks) is large enough to be informative but still the newest of the four phases.
• The 2026-08-01 05:00–07:00 CDT maintenance/deploy window is excluded from both the D10 and D11 phase aggregates by design. It no longer appears in the current 24h/7d charts (too much time has passed), and at weekly resolution in the 13-month chart it's diluted into the Jul28–Aug3 bucket, which shows a modest apdex dip (0.74, the lowest nearby week) consistent with the outage.
• Error rates across the trailing 13 months have trended down phase-over-phase (D8 0.39% → D9 0.42% → D10 0.36% → D11 0.21%), with November 2025 (0.92%) standing out as the single highest monthly error rate in the full history.

Method & caveats

• Cutovers. Drupal 9 / PHP 8 went live May 23, 2026; Drupal 10 Jul 4, 2026 — marked on every chart that spans them.
• Drupal 10 is an early read (~5 days). Its phase card uses the Jul 4–9 window directly.
• Recent spike. Jul 8 shows an isolated response jump (≈2× the prior day), coinciding with elevated reports/progress-API latency that day — an operational blip, not a migration regression.
• Source & retention. New Relic timeslice metrics (HttpDispatcher, Apdex, Errors/all), ~13-month retention — the full arc back to mid-2025 is covered. Percentiles (p95/p99) live in 8-day event data and are reported in the point-in-time reviews, not here.
• Refresh. Regenerated daily by a scheduled routine that re-queries New Relic and redeploys to this same URL.