Hosting Policy
Updated April 24, 2024
1. Background
Fast Track is committed to providing a world-class experience, which includes reliable real-time data processing and high uptime in line with our SLA. We continuously review and adjust our service configuration and infrastructure with the aim of keeping end-to-end processing of real-time data below 100ms and maintaining an uptime of 99.9%. All of this is to provide the best possible end-user experience.
Service Level Agreement (SLA): https://www.fasttrack.ai/en/sla
2. Hosting Packages
On the 1st of each month, Fast Track determines the hosting package for the previous month based on the number of registered players and the volume of transactions sent to the platform during that period.
| Size | Registered Players | TPS (Transactions per Second) | Price per Month |
|---|---|---|---|
| XS | 10,000 | 0-5 | €194 |
| S | 500,000 | 5-50 | €408 |
| M | 7,000,000 | 50-150 | €1,095 |
| L | 15,000,000 | 150-300 | €1,939 |
| XL | 25,000,000 | 300-700 | €4,704 |
| XXL | 25,000,000 + | 700+ | €8,703 |
Notes
- Each Instance is monitored separately
- A staging environment is always hosted on the XS package
- A production environment is always hosted on at least the S package
- A read replica of the database can be provided at an additional cost, quoted on request
3. Calculating Transactions Per Second (TPS)
- Frequent Monitoring: We check the number of incoming messages every 5 minutes.
- Daily Calculation: Each day, we look at the last 4032 readings (equivalent to readings taken over 14 days at 5-minute intervals). We use this data to calculate the transactions per second at peak times while ignoring the very highest peaks. This ensures that one-off influxes of transactions do not skew the results.
- Monthly Average: At the end of each month, we calculate an average of these daily readings. This monthly average is used to determine the TPS and appropriate hosting package for the client.
Rationale:
- Regular Checks: By checking every 5 minutes, we get a detailed view of the client system’s activity.
- Focusing on Realistic Peaks: By focusing on the higher end of the readings while ignoring extreme spikes, we ensure that our calculations reflect typical usage rather than rare, unusual events.
- Monthly Averaging: Averaging these readings over a month provides a fair and stable representation of the client’s usage, ensuring that occasional high loads do not disproportionately affect the hosting package.
This method ensures that our services are scaled appropriately to handle typical high loads efficiently, maintaining real-time data processing capabilities and system uptime.