Drupal Migration — Performance Monitor

Bottom line — Six weeks into Drupal 11, performance is the best of any phase to date: response time averages 603ms (down from 663ms on D10 and 679ms on the D8 baseline), the error rate is the lowest yet at 0.29%, and Apdex holds steady at 0.77 — in line with D9/D10. The 5.81M requests behind the D11 figure make this an early but already substantial read, not a thin sample. The last 24h and 7 days look healthy, with the usual overnight traffic trough and a brief error/Apdex wobble during this morning's peak load.

• D11 early read: resp 603ms / apx 0.77 / err 0.29% over 5.81M reqs since Aug 1 — ~9% faster than D10 (663ms) and ~11% faster than the D8 baseline (679ms), with the lowest error rate of all four phases. Apdex is essentially flat versus D9 (0.77) and D10 (0.78), so the win so far is in latency and errors, not perceived-satisfaction score.
• Maintenance window excluded, not hidden: the Aug 1 05:00–07:00 (America/Chicago) upgrade/maintenance window is deliberately excluded from both the D10 and D11 phase aggregates so neither version's average is skewed by the outage. That window does still register as a traffic/latency dip wherever it falls inside the rolling 24h/7-day/weekly time-series charts — it is just not counted in the phase-level D10/D11 numbers reported above.
• 24h: response time ranges from a 278ms overnight low (04:05) up to a 663ms high (02:05); the day's heaviest interval is 10:05 at ~15.7K requests with error rate climbing to its daily peak of 1.31% and Apdex easing to 0.74 — a normal load-correlated dip, not an anomaly.
• 7d: Sep 11 was the strongest day (489ms, 0.09% errors, 0.86 Apdex); Sep 13 saw a brief error uptick to 0.74% (still well within normal range) with response time otherwise stable in the high-400s to low-600s ms band all week.
• Trend read: D8→D9→D10→D11 shows response time and error rate trending down release over release (679→699→663→603ms; 0.39%→0.42%→0.36%→0.29% errors), suggesting the D11 upgrade is landing cleanly on top of the D9/D10 stabilization work rather than introducing a regression.

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.