Alerts Practical By Samson Tanimawo, PhD Published Nov 18, 2025 4 min read

Paging Load by Day-of-Week

Pages cluster by day. Patterns.

Pages cluster by day

Most production stacks page hardest on Tuesday-Thursday during business hours, with secondary peaks on deploy days.

Weekends and holidays usually have fewer pages, but the pages that fire are higher severity (no humans around to catch problems early).

Plot pages by day-of-week and hour-of-day. The pattern is your real on-call shape.

If 80% of pages happen Tuesday-Thursday 9am-6pm, a 24/7 follow-the-sun rotation is wasteful.

Split rotation: business hours primary in the right time zone, off-hours fallback rotation.

Compensate the off-hours rotation. Off-hours pages are 5-10x more painful per page.

Friday deploys page on Saturday. Don't ban Friday deploys; instead, page the deployer first, then on-call.

Tag deploys in your alerting tool. When a deploy precedes an alert by under 30 minutes, route the alert to the deployer.

Track deploy-induced page rate. Above 20% of pages tied to recent deploys means CI is missing regressions.

E-commerce: Black Friday week. Tax software: April. Banking: end of month, end of year.

Surge rotations during predicted peaks. Add a second on-call for the peak window only.

Mute non-critical alerts during the peak. Tune for high-severity catches only.

Pull 6 months of paging data. Plot by day-of-week, hour-of-day, day-of-month.

Reshape rotations to match. Bigger primary on Tuesday-Thursday, smaller fallback on weekends.

Re-evaluate every 6 months. The pattern shifts with the product.