Drupal Migration — Performance Monitor

Bottom line — Since the Aug 1 upgrade to Drupal 11, response times have improved to ~596ms average (vs. 663ms on D10 and 679ms on D8), and the error rate has fallen to its lowest of any phase at 0.29%. Apdex on D11 sits at 0.77, in line with D9 and slightly below D8's 0.79. With roughly six and a half weeks of production traffic now behind it (6.43M requests), D11 is no longer a thin sample — it already exceeds D10's total request volume and gives a solid basis for comparison. Overall the upgrade looks like a net win on speed and reliability, with apdex essentially unchanged.

• D11 vs. D10 baseline: resp time down ~10% (596ms vs 663ms) and error rate down from 0.36% to 0.29%, while apdex is flat (0.78 → 0.77) — the upgrade traded a hair of apdex for materially faster, more reliable responses.
• D11 vs. D8 baseline: resp time down ~12% (596ms vs 679ms) and error rate down from 0.39% to 0.29%; apdex is 0.02 lower (0.77 vs 0.79), a small but real trade-off worth watching as more data accumulates.
• Maintenance window excluded, not currently visible: the Aug 1 05:00–07:00 Chicago maintenance window is deliberately excluded from the D10 and D11 phase aggregates so neither version is scored on downtime. That window has since aged out of the 24h and 7-day charts (today is Sep 17), so it will not appear as a dip in the current time-series views.
• Overnight anomaly (last 24h): the 02:05 Chicago hourly bucket shows a resp-time spike to 1379ms with apdex dropping to 0.58, against low overnight traffic (~1.6K requests) — consistent with a batch/cron job or cold-cache effect rather than a user-facing incident, since the following two hours (03:05, 04:05) drop back to 291–350ms with apdex 0.87–0.91.
• Isolated error spike: the 23:05 hourly bucket shows a 6.02% error rate on very low traffic (1,313 requests) — an isolated blip, not sustained; the surrounding hours are all under 0.3%.
• 7-day trend: daily error rate ticked up on Sep 13 (0.73%) and Sep 14 (0.52%) versus a typical <0.35% for the rest of the week, while apdex stayed stable (0.75–0.86); Sep 12 shows unusually low volume (41.3K reqs vs. 150K–330K on neighboring days), worth a quick sanity check if it recurs.

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.