--- title: Hosting SLA description: "A service-level agreement covering uptime, support response, and compensation. Read the definitions and exclusions." canonical: https://penasihathosting.com/en/hosting-wiki/hosting-sla type: wiki locale: en updated: 2026-10-29 author: Willya Randika --- # Hosting SLA ## Overview - **Summary:** A service-level agreement covering uptime, support response, and compensation. Read the definitions and exclusions. - **Author:** Willya Randika ([profile](/penulis/willya-randika)) ## Article An SLA is a written promise about service levels: uptime percentage, support response time, maintenance procedures, and compensation when the promise is missed. An SLA is not a guarantee that the service will never fail. It defines how failures are measured and what happens next. ## What the uptime number means 99 percent uptime means about 7 hours 12 minutes of downtime per 30 days. 99.9 percent is about 43 minutes. 99.95 percent is about 22 minutes. 99.99 percent is about 4 minutes 19 seconds. The number only means something if down is clearly defined. Does it count only the network, or also the server, database, and application? Is scheduled maintenance excluded? Is downtime measured monthly or yearly? ## What is usually covered An SLA covers uptime and the measurement method, support response time, operating hours, maintenance windows, and compensation. Compensation is usually a service credit, not cash, with specific claim conditions. Exclusions often include force majeure, large DDoS attacks, customer misconfiguration, and third-party providers. When backups and recovery are involved, look for RTO and RPO as well. ## How to evaluate an SLA Match the uptime number to the business need. An online store may need 99.95 percent, while a blog may be fine with 99.9 percent. Check whether credits are automatic or must be claimed within a deadline. Look for compensation caps. They are often a small percentage of the monthly fee. Confirm there is a status page and incident history. Ask whether backup and restore are covered, and for how long. A good SLA also has an escalation path and emergency contact. For critical services, combine it with a [load balancer](/en/hosting-wiki/load-balancer), [anycast](/en/hosting-wiki/anycast), and offsite backups. ## Common mistakes Assuming 100 percent uptime is realistic does not help. Thinking credits replace business losses is also wrong, because credits are usually a bill discount. Without your own downtime records, a claim is hard to prove. Many SLAs require a report within 7 to 30 days. Do not wait until year-end. An SLA relates to [uptime](/en/hosting-wiki/uptime), [hosting backups](/en/hosting-wiki/hosting-backups), and [managed vs unmanaged](/en/hosting-wiki/managed-vs-unmanaged). ## FAQ No. An SLA covers service availability, not security or data integrity. Backups and security practices must be checked separately. It is usually excluded when announced according to the terms. Frequent or long maintenance is still worth questioning. Keep the ticket, start and end times, and monitoring evidence. File according to the provider's process and deadline.