Uptime SLA credits are typically 5-10% of one month's hosting fee, paid as account credit rather than cash, and only after downtime exceeds a minimum threshold — often 43+ minutes a month at a 99.9% guarantee. Real-world hosting uptime commonly clusters around 99.9-100% regardless of price, so many outage months never even qualify.
Key takeaways
- Every SLA credit checked here is account credit, not cash — applied toward future service, and explicitly forfeited if you cancel before using it.
- Most SLAs only pay out past a minimum downtime threshold per billing period — commonly tens of minutes a month, not "any downtime at all."
- Real-world uptime across budget and premium hosts alike clusters near 99.9-100%, which can mean a genuinely bad month still doesn't cross the threshold required to claim anything.
- The credit is a percentage of your hosting fee, not a reflection of what the downtime actually cost you — a $10/month shared plan's 5% credit is 50 cents, regardless of what was riding on that site being up.
- Claiming it is usually manual, not automatic — you contact support, and approval is frequently at the provider's discretion.
What four providers actually promise, in their own terms
Every provider's marketing page leads with the guarantee percentage. The credit terms — what you actually get if they miss it — live in a separate legal document most people never open.
| Provider | Guarantee | Min. downtime for credit | Credit | Form |
|---|---|---|---|---|
| Hostinger (shared) | 99.9% | Not specified | 5% of month | Account credit, manual request |
| Kinsta (entry tiers) | 99.9% | 43 min/mo | 5-10%+ of month | Account credit, capped at 1 month |
| AWS EC2 (region-level) | 99.99% | Below 99.0% | 10-100% of month | Account credit, tiered by severity |
| Bluehost (shared) | 99.99% | Not specified | 5% of month, 1 coupon cap | Account coupon, discretionary approval |
AWS's tiered structure is the outlier here in the credit's favour — a severe enough outage (region availability below 95%) pays a full month's credit, not a token percentage. But AWS's threshold for any credit at all is also stricter (below 99.99% at the top tier), and this specific SLA covers an entire AWS region's EC2 service, not an individual budget hosting account — a materially different product and audience than the other three.
The real cost of downtime vs. the credit
Take a site generating $40/hour in revenue or value, and 90 minutes of downtime in a month. That's $60 in real cost. A Hostinger or Bluehost-style 5% credit on a $10/month plan is $0.50. Kinsta's tiered credit on a $50/month plan is roughly $5. None of these come close to covering the actual loss — and that's before considering the credit isn't cash.
Bar length = dollar amount, scaled to the $60 illustrative real cost. Hosting fee figures are representative, not provider minimums.
The $40/hour figure is illustrative, not a universal number — the point is structural, not about matching your own site's economics exactly. Whatever your actual hourly value is, multiply it by realistic downtime minutes and compare that to 5-10% of your monthly hosting fee. The gap is rarely close.
Why the minimum-downtime threshold matters more than the percentage
Kinsta's published SLA pays no credit at all until downtime in a billing period passes 43 minutes at the 99.9% guarantee tier. Independent monitoring data cited for several hosts — budget and premium alike — puts real-world uptime in the 99.9-100% range across a given month. At 99.95% actual uptime, a 30-day month has about 22 minutes of downtime — under Kinsta's own 43-minute floor. A visibly bad month for your site can still land under the line that makes a credit possible at all.
Documented vs. inferred vs. unsupported
| Claim | Status | Basis |
|---|---|---|
| Kinsta's SLA requires 43+ minutes of monthly downtime at the 99.9% tier before any credit applies | Documented | Stated directly on Kinsta's own published SLA page, dated September 2025 |
| AWS EC2's region-level SLA pays tiered credits of 10%/30%/100% at stated uptime thresholds | Documented | Stated directly on AWS's own published EC2 SLA page |
| Bluehost caps its SLA credit at one coupon per month regardless of outage count, with discretionary approval | Documented | Stated directly on Bluehost's own published uptime-guarantee help page |
| Hostinger's SLA credit is 5% of the monthly charge, account-credit only | Inferred | Sourced from third-party summaries, not independently confirmed against Hostinger's own current terms-of-service page in this pass — sources also disagreed on whether a formally defined SLA document exists at all |
| Real-world hosting uptime commonly clusters around 99.9-100% regardless of price tier | Inferred | Based on third-party monitoring studies and review-site data for several named hosts — not a controlled, first-party measurement run by this site |
| Paying more for hosting always buys meaningfully better real-world uptime | Unsupported | The monitoring data found for this comparison shows budget and premium hosts landing in a similar 99.9-100% range — uptime differences at that margin come from infrastructure and operational practice, not simply price |
What this means for picking a host
- Don't weight the guarantee percentage heavily. 99.9% vs 99.99% matters less than whether the credit behind it is easy to actually qualify for and claim.
- Check the minimum-downtime threshold, not just the credit percentage. A generous-sounding credit gated behind a high threshold rarely pays out in practice.
- Treat the SLA as a mild deterrent against prolonged neglect, not insurance. It will not make you financially whole after a real outage — budget and plan accordingly, including your own backup and monitoring, regardless of which host you're on.
- If uptime is genuinely critical to your business, look for third-party, independently verified uptime monitoring data for your specific candidate hosts, rather than the SLA percentage on the pricing page.