What is an SLA?
An SLA (service level agreement) is a contract between a provider and a customer that promises a level of service, most often a monthly uptime percentage such as 99.9%, and states the remedy if it is missed, usually service credits. It is a business commitment, which is why internal SLOs are set stricter than the SLA.
Why an SLA matters
An SLA turns reliability into money and trust. Customers compare providers on it, procurement teams ask for it, and missing it costs credits and renewals.
Each extra nine cuts the allowed downtime tenfold: 99.9% allows 43.2 minutes in a 30-day month, 99.99% only 4.32 minutes, which is less than many incidents take to diagnose. Promise only what your monitoring shows you already achieve.
How to measure an SLA
allowed downtime = (1 − SLA) × period measured availability = (period − downtime) ÷ period 30-day month = 43,200 minutes
Worked example
Your contract promises 99.95% monthly uptime. In a 30-day month (43,200 minutes) the service was down for 50 minutes in total.
- Allowed downtime = 0.0005 × 43,200 = 21.6 minutes.
- Measured availability = (43,200 − 50) ÷ 43,200 = 99.88%.
- 99.88% is below 99.95%, so the SLA was missed and the credit terms in the contract apply.
Result: The month would have passed a 99.8% SLA (86.4 minutes allowed). The gap between what you promise and what you deliver is a business decision, so base it on measured history.
Common mistakes
- An SLA equal to the SLO. If the internal target is the same as the contract, you only learn you are in trouble when you already owe credits.
- Undefined measurement. Write down how availability is measured: from where, how often, what counts as down, and whether maintenance windows are excluded.
- Ignoring dependencies. Your SLA cannot sensibly exceed what your cloud, DNS and payment providers give you, unless you have redundancy.
- Calendar month confusion. A 31-day month allows more minutes than a 28-day one at the same percentage. Say which period you use.
How Vigil handles an SLA
- The free SLA and uptime calculator converts any percentage into downtime per day, week, month and year and works backwards from minutes of downtime; common targets such as 99.9% and 99.95% have their own pages.
- Set SLO targets on monitors and print the SLA report for a customer or an auditor; maintenance windows show on status pages.
- Every failure is confirmed by a second check before it counts, so a single network blip does not eat into measured availability.
SLA (service level agreement): frequently asked questions
What is the difference between an SLA, an SLO and an SLI?
The SLI is the measurement (for example, the share of successful requests), the SLO is your internal target for it, and the SLA is the external contract with consequences if a promised level is missed.
How much downtime does 99.9% allow?
43.2 minutes in a 30-day month and about 8.76 hours in a 365-day year.
Do SLAs include planned maintenance?
Many exclude announced maintenance windows and some do not. The contract must say so, and the status page should show the windows.