Even during peak, the invoice must go out on time. A failed queue risks penalties for late tax documents.
Christmas peak with no outage: queue, rate-limit and 90% throughput
Between Black Friday and 24 December, Czech email and SMS run 5× normal volume. Standard queues collapse, OTP messages arrive late, transactional confirmations wait for hours. 4notify ships a dedicated peak mode: weighted priority queue (OTP > transactional > marketing), dynamic per-operator rate-limit and IP pre-warming from 1 November.
In 2024 a major Czech e-shop sent its Christmas newsletter at 9:00 and choked its own transactional SMS queue. Result: 600,000 customers missed the „parcel arriving today“ alert and support fielded 12,000 calls. Without queue prioritisation, the Christmas disaster is a matter of when.
Recommendation not to send marketing outside this window; 4notify auto-defers non-critical messages outside the window to the next morning.
An operational outage is an incident; a defensible queue design (OTP prioritisation) reduces harm risk.
Three queue tiers
OTP and security alerts (weight 100) > transactional (weight 50) > marketing (weight 1). Under load we degrade marketing first.
Dynamic rate-limit
Per-operator throughput shifts in real time based on ACK rate. If O2 RTT > 3 s, we throttle and shift to a backup channel.
IP pre-warm from 1 November
We gradually grow IP reputation so Seznam and Gmail don't drop deliverability under the sudden 5× volume spike on Black Friday.
Degradation mode
Above 90% SLA throughput we auto-shift non-critical messages to a delay queue (max 6 h) and notify the admin.
json
// Queue weights configuration (4notify > Settings > Peak Mode):
{
"peak_mode": {
"enabled_window": { "from": "2026-11-24", "to": "2026-12-31" },
"weights": {
"otp": 100,
"security": 100,
"transactional": 50,
"marketing": 1,
"newsletter": 1
},
"degradation_at_load_pct": 90,
"ip_prewarm_start": "2026-11-01",
"communication_window": { "from": "07:00", "to": "21:00" }
}
}Hello, your environment is ready for the Christmas peak. Peak mode runs 24 Nov to 31 Dec, OTP priority 100, marketing priority 1. Contact on-call via #vanocni-podpora if needed.
- Enable peak_mode from 24 Nov to 31 Dec
- Schedule IP pre-warm from 1 Nov
- Audit queue weights (OTP must be 100)
- Set alert at 90% SLA throughput
- Stage rehearsal at 3× normal volume
- Enable on-call rotation via Slack / PagerDuty
Weighted priority queue + IP pre-warm + graceful degradation — keeps OTP latency under 3 s through Black Friday and Christmas peak.
How much does peak mode cost?
Peak mode is free on Growth and Scale plans. Above the contracted volume we bill variably per price list, but OTP and transactional stay at the base rate.
What if Seznam still drops deliverability?
4notify has a dedicated standby IP pre-warmed; switch takes < 60 s and Seznam sees it as known. Loss-of-delivery risk is below 0.1%.
How do you handle the Christmas carrier-error spike?
The EXCEPTION queue grows 4× in November. 4notify ships a real-time dashboard with suggested actions („add 2 support agents“, „enable overflow pickups“).
Start free
14 days, no card. Czech-language support.