Drupal Migration — Performance Monitor

Bottom line — Drupal 11's first ~4 days on live look healthy: response time (625ms) and error rate (0.28%) are the best of all four phases, beating both the immediately-prior D10 phase (663ms, 0.36%) and the original D8 baseline (679ms, 0.39%). Apdex sits at 0.77, essentially flat with D10 and D9 but still a touch below the D8 baseline of 0.79. The 2026-08-01 05:00–07:00 CT maintenance window is excluded from the D10/D11 phase aggregates, so neither version is penalized for the upgrade downtime.

• D11 early read: resp 625ms / err 0.28% / apdex 0.77, on 462.67K requests — a real but young sample (~4 days vs weeks of history for D8–D10). Directionally better than D10 on speed and errors so far; worth re-checking once D11 accumulates a comparable multi-week window before calling it a durable win.
• Error rate trend: improving phase over phase — 0.39% (D8) → 0.42% (D9) → 0.36% (D10) → 0.28% (D11) — D11 continues that downward trajectory rather than reversing it.
• Maintenance dip visible in raw series: the Aug 1 daily bucket in the 7-day view shows ~42K requests versus ~90–190K on neighboring days — that's the 05:00–07:00 CT upgrade window. It's intentionally excluded from both phase aggregates but will keep showing as a dip in the 24h/7d charts.
• Overnight response-time spike: the 24h view shows an isolated hour (03:09 CT) at 1114ms response time with apdex dropping to 0.70 — the single worst hour in the last day, with no recurrence in adjacent hours. Worth a quick check if it repeats tonight.
• Apdex still shy of D8: across D9–D11 apdex has hovered at 0.77–0.78, versus 0.79 on the original D8 baseline — a small but persistent gap that the D11 upgrade hasn't closed yet.

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.