gPdf health

24h figures sourced from Cloudflare Analytics over real customer traffic.

Two data sources, one window. Status indicator and 24-hour heatmap come from a synthetic probe — a 1-page label render submitted via POST /api/v1/pdf/render every 5 minutes. The heatmap renders one coloured cell per 10-minute bucket across the trailing 24 hours, so each bucket pairs two consecutive probes. A bucket flips red only when every probe in it fails AND it contains at least two probes — i.e. a confirmed back-to-back outage. Single-probe failures or buckets with a recovery probe stay green; their counts surface on hover. Faint grey = no probe in the bucket (only seen if the cron worker missed a window). p50 / p99 / 24h uptime / requests come from Cloudflare Analytics over the trailing 24 hours of real customer traffic — the same number the Hero metric strip on the homepage reads, so values stay coherent between the two surfaces. Source: functions/api/health.js, functions/api/_health-cron.js, functions/_lib/cf-analytics.js.