Drupal Migration — Performance Monitor

Bottom line — Spectre_5 (live) is stable after the Aug 1 Drupal 11 upgrade: response time (673ms) and error rate (0.27%) in the D11 window are both in line with or better than the D8/D9/D10 baselines, though this is only a ~5-day, ~694K-request early read against phases that each span weeks and millions of requests. Apdex has drifted down slightly across every phase (0.79 → 0.77 → 0.78 → 0.75 D8→D11), a gradual trend that predates D11 and isn't something the upgrade caused on its own. The most recent 24h shows response time climbing toward 900–1170ms in the last few hours, worth a follow-up look if it persists.

• D11 early read: resp 673ms and err 0.27% both land between the D8 baseline (679ms / 0.39%) and D10 (664ms / 0.36%) — no regression signal yet, but the ~694K-request sample is thin next to D8/D9/D10's multi-million-request totals, so treat it as provisional.
• Apdex softening: score has eased from 0.79 (D8) to 0.75 (D11), the lowest of the four phases; not dramatic, but worth watching as D11 accumulates more traffic.
• Aug 1 maintenance dip: the daily view shows Aug 1 volume dropping to ~41.8K requests versus ~130–190K on adjacent days — that's the 05:00–07:00 Chicago maintenance window bleeding into the day-level bucket. It's visible in the 24h/7d/full time series as expected, but is excluded from both the D10 and D11 phase aggregates so neither version is scored on the downtime.
• Last 24h response-time creep: hourly response time has risen from the low 700s to 860–1170ms over the last several hours with error rate also ticking up to 0.19–0.37% — still within normal range for this app but a pattern to keep an eye on for the next refresh.
• Error rate trend: error rate has trended down phase-over-phase (0.39% → 0.42% → 0.36% → 0.27%), with D11's 0.27% the lowest of the four — a positive early signal, again on a small sample.

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.