Cato uses a simple approach: if a Covered Instance experiences qualifying Unavailability, the customer receives a Service Credit equal to 5× the downtime, subject to the SLA’s billing-cycle cap.
For example, if a server is unavailable for 3 hours, the credit is calculated using 15 hours of service.
There is no minimum downtime threshold before credits begin.
A 100% commitment does not mean we expect infrastructure to be incapable of failing. It means there is no amount of qualifying downtime that we treat as acceptable before the SLA begins providing a remedy. If a Covered Instance has qualifying Unavailability, the SLA credit applies.
Many traditional datacenter SLAs use monthly availability tiers. For example:
| Monthly Availability | Service Credit |
|---|---|
| 99.0%–99.5% | 5% |
| 98.0%–99.0% | 10% |
| 95.0%–98.0% | 20% |
| Below 95.0% | 30% |
Those models create thresholds. A customer can experience downtime without receiving any credit at all, and once the maximum tier is reached, additional downtime may produce no additional remedy.
Cato instead uses a continuous calculation: each qualifying period of Unavailability earns a 5× credit, up to the Fees for the affected Covered Instance for that billing cycle.
For a server with $1,000 in Fees for a 730-hour billing cycle:
| Downtime |
Traditional Tier Credit |
Cato Credit |
|---|---|---|
| 1 hour | $0 | $6.85 |
| 7 hours | $50 | $47.95 |
| 14 hours | $100 | $95.89 |
| 35 hours | $200 | $239.73 |
| 48 hours | $300 | $328.77 |
| 100 hours | $300 | $684.93 |
| 200 hours | $300 | $1,000 |
The exact credit depends on the Fees for the affected Covered Instance and the duration of qualifying Unavailability.
It is simpler and avoids arbitrary cutoff points.
There is no question about whether an outage happened to cross the 99.5%, 99%, or 95% threshold. Qualifying downtime is measured, the applicable Fees are calculated, and the 5× multiplier is applied.
The result is an SLA where the remedy grows with the outage instead of jumping between tiers or stopping at a relatively low percentage.
