Drupal Migration — Performance Monitor

Bottom line — Spectre_5 is stable over the last 24 hours and week, and the Drupal 11 phase (since Aug 1) is now running measurably better than every prior baseline: average response time is down to 594ms versus 664ms on D10 and 679–699ms on D8/D9, with the lowest error rate of any phase (0.29%) and an apdex score (0.77) essentially level with D9/D10. This is no longer a thin early read — D11 has accumulated roughly 6.66M requests over its ~7 weeks in production, comparable in scale to the other phases, so the improvement looks like a real, sustained gain rather than a small-sample fluke.

• D11 vs baselines: response time improved ~11% versus D10 and ~13–15% versus D8/D9, while the error rate roughly halved from the D9 peak (0.42% → 0.29%); apdex held steady rather than improving, suggesting the win is mostly on raw latency and reliability rather than user-perceived responsiveness thresholds.
• Overnight spike, last 24h: a sharp anomaly at 02:06 CDT (resp ~1961ms, apdex dropped to 0.48, ~1.2K requests in that hour) with a smaller precursor at 01:06 CDT (resp 745ms, err 1.42%) — both in the lowest-traffic overnight window, consistent with an isolated batch job or cache-warm event rather than a load-driven issue.
• 7-day trend: daily response time ranged 457–618ms with no sustained degradation; Sep 13 shows a mild error-rate bump (0.73%) that did not recur on subsequent days.
• Aug 1 maintenance window: the 05:00–07:00 America/Chicago deploy/maintenance window is deliberately excluded from both the D10 and D11 phase aggregates so neither version is scored on downtime. That window is now roughly seven weeks in the past, so it no longer falls inside the current 24-hour or 7-day time-series charts (both of which only cover Sep 11 onward) — it would only reappear if those charts are widened to include early August.
• Weekly/monthly view: the weekly series shows response time trending down through August (Aug 7: 615ms → Aug 21: 557ms) into the D11 period, reinforcing that the post-upgrade improvement has been consistent rather than a one-off week.

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.