Drupal Migration — Performance Monitor

Bottom line — Drupal 11 is off to a strong early start: in the ~3 days since the Aug 1 upgrade it is running the fastest (625 ms avg response) and cleanest (0.26% error rate) of the four phases tracked, on par with D9/D10 for Apdex (0.77). D8/D9/D10 remain stable in the 660–700 ms band, with D9 the weakest phase on both error rate (0.42%) and response time (699 ms). No regression is visible in the live 24h/7d series post-upgrade.

• D11 is a thin, early sample — ~326K requests so far vs. 3.1–6.8M for the other phases; today's favorable numbers (625 ms, 0.26% err) should be treated as a promising first read, not a settled baseline, until D11 accumulates comparable volume.
• D11 vs. D10/D8 baselines — response time is ~6% faster than D10 (663 ms) and ~8% faster than the D8 baseline (679 ms); error rate is roughly a third lower than D10's 0.36%. Apdex is essentially flat across D9/D10/D11 (0.77–0.78), so the gains are in latency/errors, not perceived satisfaction score.
• Maintenance window visible in traffic, excluded from scoring — the Aug 1 05:00–07:00 Chicago deploy/maintenance window is deliberately excluded from both the D10 and D11 phase aggregates, but it still shows up as a clear dip in the 7‑day series: Aug 01 logged only ~41.8K requests for the day vs. 87K–182K on surrounding days.
• Overnight latency blip, low confidence — the 24h series shows a spike to 1172 ms at 03:09 CDT, but on very light overnight traffic (1589 requests) — consistent with small-sample noise rather than a systemic issue; the surrounding hours stay in the 400–700 ms range.
• Error rate steady, no post-upgrade anomaly — hourly error rate over the last 24h stays mostly under 0.3%, with an isolated 0.63% blip at 01:09 CDT again coinciding with a low-traffic hour; nothing in the post-upgrade window suggests elevated errors from the D11 code change.

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.