The verdict in three sentences
A ticket sale opening packs the equivalent of several weeks of normal traffic into 10 minutes, and that is exactly when an untested platform falls over. A load testing campaign with k6 or Gatling, a virtual waiting room and tuned autoscaling costs, as a 2026 order of magnitude in London, £4,300 to £13,000 (5,000 to 15,000 EUR) over 3 to 5 weeks. Compared with the revenue of a sales opening (often several hundred thousand pounds in one hour), it is the cheapest insurance in the project.
What to measure before launch day
The CTO of a festival promoter does not want a theoretical report: they want to know at how many users checkout slows down, at what threshold the database saturates and how many servers must be running at 9:59 am. Tests split into distinct scenarios, each with a numeric target.
| Test type | Goal | Typical load (40,000-connection festival) | Run time |
|---|---|---|---|
| Smoke test | Validate scripts and environment | 50 virtual users | 10 to 15 min |
| Nominal load test | Hold the expected load | 15,000 active users, 400 purchases/min | 30 to 60 min |
| Spike test | Absorb the sales opening | 0 to 40,000 connections in 2 min | 15 to 20 min |
| Stress test | Find the breaking point | Ramp up to 60,000 connections | 30 to 45 min |
| Soak test | Detect memory leaks and drift | 8,000 constant users | 4 to 8 h |
| Waiting room test | Check admission order and rate | 40,000 queued, 500 admitted/min | 30 min |
The metrics are few but non-negotiable: 95th percentile response time under 800 ms on seat selection pages, under 2 s on basket validation, error rate below 0.5%, and zero overselling (two people must never get the same numbered seat). That last point is tested with concurrent scenarios on the same seats, not just with volume.
What a campaign costs in 2026
The prices below are 2026 orders of magnitude observed on the London market, excluding the ticketing licence. They depend mostly on the number of journeys to script and the maturity of the staging environment.
| Item | 2026 range | Comment |
|---|---|---|
| Architecture audit and test plan | £1,300 to £2,600 | Identify bottlenecks: database, cache, payment, sessions |
| Writing k6 or Gatling scripts | £1,700 to £4,300 | 4 to 8 journeys: browsing, seat choice, basket, payment |
| Execution and analysis (3 to 4 rounds) | £1,300 to £3,500 | Each round followed by fixes and a new run |
| Virtual waiting room setup | £0 to £2,600 | Managed service or in-house build |
| Autoscaling and CDN cache tuning | £430 to £1,700 | Scaling rules, cache pre-warming |
| Peak infrastructure (runs and launch day) | £260 to £1,700 | Cloud load generators and extra servers |
| Total campaign | £4,300 to £13,000 | 3 to 5 calendar weeks |
k6 (open source, JavaScript scripts) suits web teams well, while Gatling (Scala or Java, very readable reports) is favoured for complex simulations. Both generate tens of thousands of virtual users from a few cloud machines. A managed virtual waiting room is often billed by event volume, from a few hundred to a few thousand pounds per season: compare that with an in-house build, which requires careful handling of admission tokens.
A realistic schedule
| Week | Deliverable | Decision expected |
|---|---|---|
| W1 | Audit, numeric targets, smoke scripts | Validate thresholds (p95, error rate) |
| W2 | Full scripts, first nominal load run | Prioritise bottlenecks found |
| W3 | Fixes, spike run, waiting room tuning | Set admission rate |
| W4 | Stress and soak tests | Size launch day servers |
| W5 | Dress rehearsal, runbook, on-call | Go or postpone the opening |
What derails schedules is rarely the test tool: it is a staging environment that does not resemble production, or a payment provider that refuses runs on its sandbox beyond a few requests per second. Plan a payment simulator and warn suppliers (CDN, host, PSP) of the test dates.
Mini case study
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Oliver, CTO of a festival promoter in London, is preparing the opening of a 3-day festival: 45,000 tickets at £76 on average. He estimates 30% of sales happen in the first hour, that is 13,500 tickets and about £1.03M. A 20-minute outage would, by his cautious estimate, lose 15% of those sales to resale sites or abandonment, roughly £154,000 of revenue at risk, before counting the reputational cost on social media. He commissions a campaign at £9,400 (k6 scripts, 4 rounds, managed waiting room, autoscaling) plus £1,000 of peak infrastructure. The second round reveals a database lock on seat reservation that saturated at 600 purchases per minute: once fixed, the system holds 1,100 purchases per minute. The total cost represents about 1% of first-hour revenue.
FAQ
How many virtual users should we simulate?
Aim for at least 1.5 times the expected peak: for 40,000 announced connections, a stress test at 60,000 is reasonable. Marketing estimates are often exceeded by 20 to 50% for headline acts.
k6 or Gatling, which one?
k6 is simpler for a JavaScript team and plugs easily into CI. Gatling offers detailed reports and suits complex scenarios; both are free in their open source versions, only the load generation infrastructure is paid (often £40 to £260 per round).
Is a virtual waiting room essential?
Above 10,000 simultaneous connections, it is the most effective protection: it smooths arrivals to 300 or 500 admissions per minute. It also protects perceived fairness, which reduces complaints.
Can we test directly in production?
Yes, outside sales hours and with simulated payment, but only after staging runs. Warn the host and CDN 48 h ahead, otherwise the run may be blocked as an attack.
Do we need to retest every season?
A 1 to 2 day regression run (£1,300 to £2,600) is enough if the architecture has not changed. A new payment or seating feature calls for a fuller campaign.
Let's scope your project. Tell us your expected peak, your stack and your opening date: we price the test campaign, the waiting room and the sizing. Detailed quote within 48 h. WhatsApp +221 77 596 93 33.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.
