SLA (service level agreement)
An SLA (service level agreement) is a promise to customers about the level of service, such as 99.9% monthly uptime, and what they receive if it is not met.
A typical SLA names the metric, the target, the period it is measured over, what is excluded (announced maintenance, customer-caused problems, force majeure) and the remedy, usually a service credit worth part of that period's fee.
The details matter as much as the number. 99.9% per month and 99.9% per year allow the same total downtime, but the monthly one does not let a single long outage be averaged away. Credits often have to be claimed within a deadline.
Teams usually keep an internal SLO stricter than the SLA, so they notice and fix trouble before the contract is breached.
In OpenUptime
Uptime history in OpenUptime is kept per monitor and shown on status pages, which gives customers and you the same record to check an SLA against.
Questions
What is a good SLA for a SaaS product?
99.9% a month is common. Promise only what your own monitoring shows you already meet, with some margin, because the SLA is a cost when it is missed.
What is the difference between an SLA and an SLO?
The SLA is an external promise with consequences; the SLO is an internal target. The SLO is set stricter so it gives warning before the SLA is at risk.
Related terms
Monitor it with OpenUptime
Checks from one or several regions, alerts to the tools you use, and a status page for your customers. Free to start, open source under the MIT license.
SLA（服务等级协议）
SLA（服务等级协议）是对客户的服务水平承诺，例如每月 99.9% 的可用率，以及未达成时客户能得到什么补偿。
一份典型的 SLA 会写明：衡量指标、目标值、统计周期、排除项（提前公告的维护、客户自身原因、不可抗力等），以及补偿方式，通常是当期费用一部分的服务抵扣。
细节和数字一样重要。按月 99.9% 和按年 99.9% 允许的停机总量相同，但按月计算不能把一次长时间故障平摊掉。抵扣通常也需要在规定期限内申请。
团队通常会设定比 SLA 更严格的内部 SLO，这样能在违约之前发现并解决问题。
在 OpenUptime 中
OpenUptime 按监控保存可用率历史，并在状态页上展示，让你和客户可以对照同一份记录核对 SLA。
常见问题
SaaS 产品的 SLA 定多少合适？
按月 99.9% 比较常见。只承诺你的监控数据显示已经能稳定达到、并留有余量的水平，因为达不到 SLA 是要付出代价的。
SLA 和 SLO 有什么区别？
SLA 是对外的承诺，违约要承担后果；SLO 是内部目标。SLO 设得更严格，以便在 SLA 面临风险之前发出预警。