Drupal Migration — Performance Monitor

Bottom line — Since going live on Drupal 11 (Aug 1), the site has posted its best response time and lowest error rate of any phase to date — 596ms avg / 0.29% errors, down from Drupal 10's 664ms / 0.36% and Drupal 8's 679ms / 0.39% — on comparable traffic volume (6.32M requests over ~46 days, in line with D8/D9/D10). Apdex on D11 sits at 0.77, essentially flat with D9 and a touch below D10 (0.78) and D8 (0.79), so the latency and error gains haven't yet fully translated into a higher user-satisfaction score. Overall the D11 upgrade looks like a clean win on the metrics that matter most (speed, errors) with no regression signal six-plus weeks in.

• D11 vs. baselines: Drupal 11's 596ms average response is the fastest of the four phases (10-13% quicker than D8/D9/D10), and its 0.29% error rate is the lowest observed — both good early signs, though the sample now spans ~46 days so this is a solid read, not a thin one.
• Apdex didn't move with latency: despite the response-time and error-rate improvement, D11's Apdex (0.77) trails D10 (0.78) and D8 (0.79) slightly. Worth a periodic re-check as more data accumulates rather than treating latency and error gains as the full picture.
• Overnight traffic spike, 24h view: requests jumped roughly 4-6x between 22:06 and 01:06 (up to ~68.7K in one hour) while response time dropped to 270-380ms and Apdex hit 0.95-0.99 — consistent with cached/crawler traffic rather than a problem.
• Daytime latency peak: the 09:06 hour bucket shows the day's slowest response (761ms) and lowest Apdex (0.71-0.72) in the 24h window, coinciding with a return to normal (non-cached) traffic levels.
• 7-day error blip: Sep 13 shows the week's highest error rate (0.73%, vs. a 0.09-0.52% range on other days) with otherwise-normal traffic and response time — an isolated day worth a closer look if it's unfamiliar.
• Maintenance window note: the Aug 1, 2026 05:00–07:00 Chicago cutover to Drupal 11 is excluded from both the D10 and D11 phase aggregates so neither version is scored on downtime. That window is now over six weeks in the past, so it no longer appears in the 24h/7d series above; it shows only as a minor, heavily-smoothed dip in the long-range weekly ("full") view around Jul 29-Aug 5.

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.