What is an uptime SLA?
An uptime SLA is a provider's written promise of how much of each month its service will answer, and what it gives back when it does not.
Runtime promises paid accounts 99% API uptime every calendar month, measured from outside its servers, and pays 10% of a short month's charges back as credit automatically (Uptime Promise). Nobody has to file a claim.
The three parts of an SLA
- The figure. A share of the month, such as 99% or 99.9%. Each extra nine cuts the downtime allowed by ten.
- The measurement. Who checks, how often and from where. A check run from outside the provider's own network sees what customers see.
- The remedy. What a missed month earns, and whether you must ask for it. Most remedies are service credit rather than cash.
A figure without a stated measurement and remedy is a target, not a promise.
What the nines allow in a 30-day month
| Uptime | Downtime allowed |
|---|---|
| 99% | 7 hours 12 minutes |
| 99.9% | 43 minutes |
| 99.99% | 4 minutes |
Why it matters for agent workloads
An agent that runs unattended calls the sandbox API hundreds of times an hour. When that API is down, the agent's work stops with it, often at night with no one watching. An SLA tells you how often to expect that, and whether the provider pays for the time it costs you.
On Runtime
- Figure: 99% of each UTC calendar month, which leaves 7 hours and 12 minutes of downtime in a 30-day month.
- Measurement: a check from outside Runtime's servers calls the API every two minutes, and its results are published hour by hour on the status page. A real sandbox is also started and stopped from outside every ten minutes.
- Remedy: a month below the promise pays every paid account 10% of what it was charged that month, as service credit, in the first seven days of the next month. Owners are emailed when it lands.
Related
Sources
Checked 27 September 2026.