Support SLA Tracker.

Build a support SLA tracker for the support lead at Copperkite, a fictional B2B billing platform with a five person support team (Ana, Dev, Marisol, Kenji, Ruth). SLA targets by priority: Urgent gets a 30 minute first response and 4 hour resolution; High gets 1 hour and 8 hours; Normal gets 4 hours and 24 hours; Low gets 8 hours and 72 hours. Layout: two tabs, Queue (default) and Trends. Queue tab: a banner strip at the very top stating how many open tickets are within 2 hours of a resolution breach (seed it so it reads 5 tickets at risk) with red emphasis. Below, the open ticket table sorted by time to breach ascending: ticket id, subject, customer, priority badge, category, assignee, age, first response status (Met or Pending, with Pending past target shown red), and time remaining before resolution breach, colored red under 1 hour or breached, amber under 4 hours, green otherwise. Filters for priority, category, and assignee. A member only New ticket button opens a dialog capturing subject, customer, priority, category, and assignee, and members can mark first response sent or mark resolved from a row menu. Visitors must get a fully readable queue with none of that. Trends tab: four stat cards: breach rate this week (about 6 percent), median first response (about 42 minutes), median resolution (about 7 hours), tickets this week (about 38). Then charts computed from history: a line of weekly breach rate over the last 8 weeks, a bar chart of median first response by priority shown against its target, a stacked bar of weekly volume by category, and a horizontal bar of total breaches by category. Data to seed: about 175 resolved tickets spread over the last 8 weeks and 18 open tickets, across all four priorities and five categories: Billing, Onboarding, Bug report, How-to, Account. Give resolved tickets realistic first response and resolution durations, mostly inside target. Plant this story: Billing accounts for around 60 percent of all breaches so the category breach chart makes it unmistakable, and the near-breach open tickets cluster in Billing too, so the queue and the trends tell the same story. Subjects should read like a real billing product (Proration looks wrong on upgrade, Invoice PDF missing PO number). Behavior: breach math, medians, and rates are always computed from ticket timestamps and targets, never stored. Visual identity: a slate blue accent on a cool light base for tabs, primary actions, and chart emphasis, with red reserved strictly for breach states.