Drupal Migration — Performance Monitor

Bottom line — Spectre_5 is healthy over the last 24 hours and 7 days, with response times mostly in the 460–850ms range (apdex 0.71–0.81) outside one brief overnight spike. Just over two months into the Drupal 11 upgrade, the new platform is now outperforming both the Drupal 10 window it replaced and the original Drupal 8 baseline on every core metric: faster average response (591ms vs 664ms/679ms) and a lower error rate (0.21% vs 0.36%/0.39%), with apdex holding steady at 0.77.

• Overnight 24h anomaly: at 02:09 CT response time spiked to ~2,085ms and apdex dropped to 0.65 during a low-traffic hour (~2,700 reqs); it fully recovered within the next hour with no meaningful error-rate impact.
• Drupal 11 early read: D11 has logged ~9.47M requests since the Aug 1 cutover — a sizeable sample at this point, not a thin one — and already shows the best response time (591ms) and lowest error rate (0.21%) of all four phases, with apdex (0.77) in line with D9/D10.
• D10 vs D8 baseline: the now-closed D10 window (Jul 4 – Aug 1, cut before the maintenance gap) averaged 664ms/0.36% errors — a modest improvement on D8's 679ms/0.39%, but still behind D11's post-upgrade numbers.
• Maintenance window: the Aug 1 05:00–07:00 CT deploy/maintenance window is excluded from the D10 and D11 phase aggregates so neither version is scored on downtime. It no longer falls inside the current 24h/7d charts, and even the weekly-resolution 'full' series shows no visible dip for that week (resp ~650ms) since a ~28-minute gap is smoothed out at weekly granularity.
• 7-day volume dip: Oct 3 shows a request-volume trough (~50K vs 150K–220K on surrounding days) with no corresponding rise in latency or errors — consistent with a traffic lull rather than an incident.

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.