A Service Provider Agreement (SLA) is a contract outlining the quality of service (S&A) a provider promises to you. It clearly states three things in black and white: what the provider will provide, to what extent it will provide it, and how it will compensate if it fails to deliver. What happens if you encounter a failure after purchasing a server that doesn't require registration? How much compensation will you receive? What situations are excluded from compensation?—The answers to these questions are all in the SLA.
Service Availability—What Does That String of "9s" Really Mean?
When you open any cloud service provider's SLA, the first thing you'll see is that string of numbers: 99.9%, 99.95%, 99.99%... This is the most crucial metric in the SLA—Service Availability (Uptime).
The availability percentage determines how long your server, which doesn't require registration, is allowed to be downtime within a service cycle (usually a calendar month). Converted to specific timeframes:
- 99.9% (three nines): Approximately 43 minutes of downtime allowed per month
- 99.95%: Approximately 21.6 minutes of downtime allowed per month
- 99.975%: Approximately 10.8 minutes of downtime allowed per month
- 99.99% (four nines): Approximately 4.3 minutes of downtime allowed per month
99.9% and 99.99% differ by only one nine, but the allowed downtime per month differs by nearly 10 times. If your business is e-commerce, payment, or any website sensitive to online time, 99.9% means potentially 43 minutes of business interruption per month—the lost orders and customers during those 43 minutes could far exceed the savings on the server rental fee for the free registration service.
What should you pay attention to when reviewing an SLA?
First, confirm the calculation method for availability—is it calculated from the "rack switch port" or the "IDC core router"? The latter only covers the internal network of the data center, while the former covers the entire link from you to the carrier.
Second, confirm whether planned maintenance is included in downtime. Some SLAs exclude "planned maintenance" from availability calculations and have no limit on maintenance duration—meaning service providers can legally perform "maintenance" for a few hours each month without triggering compensation.
Third, confirm whether availability is based on a single instance or multiple availability zones. A single instance commitment of 99.975% and a multiple availability zone commitment of 99.995% involve completely different architectural investments and stability guarantees.
Compensation Mechanism—How much and how will you be compensated if something goes wrong?
What if availability doesn't meet standards? Compensation terms are the second most important part of an SLA.
Most service providers compensate with vouchers or service time deductions, not cash. In other words, the service provider doesn't compensate you with money, but rather with "a deduction that can be used when you buy a server without registration next time."
The compensation amount is usually linked to the actual availability level achieved. For example:
- Availability between 99% and 99.9% → Compensation of 10%-30% of the monthly service fee
- Availability below 95% → Compensation of 30% or even higher of the monthly service fee
What should you pay attention to when reviewing an SLA?
First, confirm the compensation cap. Most standard SLAs have a compensation cap of "1-3 times the monthly service fee." If your business is interrupted for 8 hours and you lose 100,000 yuan, but your monthly hosting fee is only 3,000 yuan, even with 3 times the compensation, you'll only receive 9,000 yuan—a huge difference from the actual loss.
Second, confirm the compensation trigger conditions. Is compensation triggered by "any single SLA failure," or only after "two consecutive failures"? The stricter the conditions, the more difficult it is to obtain compensation.
Third, confirm whether compensation is automatically triggered or requires user application. Some SLAs require users to submit a compensation application after the failure occurs; applications submitted after the deadline will not be accepted.
Disclaimer Clauses—What situations are not covered?
This is the most easily overlooked part of an SLA, yet the part that most likely prevents you from receiving compensation.
Almost all SLAs list several exclusions—in which case, even if service is interrupted, the service provider is not liable for compensation.
Common exclusions include:
- Force Majeure: Natural disasters, war, government actions, etc.
- User-related problems: User configuration errors, operational mistakes, software problems, etc.
- Third-party factors: Carrier network failures, upstream supplier problems, etc.
- DDoS attacks: Service unavailability caused by large-scale DDoS attacks
- IP blocking: Unavailability caused by IP address blocking or filtering by a third party
What should you pay attention to when reviewing an SLA?
Focus on whether the boundaries of the exclusions are too broad. Some SLAs include "network carrier failures" in the exclusions—this means that if the service provider's own upstream network goes down, your service is interrupted, but you will not receive any compensation.
Also, confirm whether planned maintenance is notified in advance. A reasonable SLA will require the service provider to give 72 hours or even 7 days' notice for planned maintenance and limit the cumulative maintenance time per month/year. Response and Repair Time – How Long Does It Take for Someone to Address a Problem?
If a server without registration goes down, how long will it take for the service provider to respond? How long will it take to fix it? These two metrics are usually agreed upon separately in the SLA.
Response Time: The time from when you submit a support ticket or request for assistance to the first response from the service provider's customer service personnel. Industry Standard: Response within 30 minutes for general faults, and within 15 minutes for emergency faults.
Repair Time: The time from fault confirmation to service restoration. This usually depends on the specific type of fault – hardware faults may take several hours to replace parts, while network faults may only require switching to a backup line within minutes.
Monitoring and Reporting – Who Proves “Something Happened”?
SLA execution relies heavily on monitoring and data. Who monitors whether the service meets standards? What tools are used for monitoring? Who controls the monitoring data?
The answers to these questions determine whether you have evidence to claim compensation when you believe the service provider has failed to meet standards.
What should you pay attention to when reviewing an SLA?
First, confirm the monitoring party. Is the fault diagnosis solely monitored and determined by the service provider, or are customers allowed to provide third-party monitoring evidence (such as Pingdom, Datadog, etc.)? The latter is more advantageous for you.
Second, confirm whether a real-time monitoring dashboard or monthly availability report is provided. Transparent data allows you to understand the true status of the service at any time.
Maintenance Window – When will there be "downtime"?
Servers requiring no registration need regular maintenance – system upgrades, hardware replacements, network adjustments – these maintenance will cause temporary service interruptions.
What should you pay attention to when reviewing an SLA?
Confirm whether planned maintenance is included in availability calculations. If so, the service provider's availability commitment is compromised – they can promise 99.9% availability while spending several hours each month on maintenance, and this maintenance time doesn't count as "unavailable".
Confirm whether maintenance is notified in advance and how much advance notice is required. A reasonable SLA will require at least 72 hours' notice, giving you time to prepare.
Data Backup and Recovery – What if data is lost?
For servers storing important data, the data protection clauses in the SLA are also worth paying attention to. RTO (Recovery Time Objective): The maximum time required for the service to return to an available state after a failure.
RPO (Recovery Point Objective): The maximum acceptable window of data loss. For example, an RPO of 1 hour means that up to 1 hour of data could be lost.
What should you pay attention to when reviewing an SLA?
Confirm whether the service provider offers automatic backup functionality, the backup frequency, and the data recovery process and timeframes.
Dispute Resolution – What if you can't reach an agreement?
If there is a dispute regarding the implementation of the SLA, the SLA will usually stipulate the dispute resolution method.
What should you pay attention to when reviewing an SLA?
Confirm the jurisdiction and arbitration institution. If the service provider is in Hong Kong and you are in mainland China, and the dispute resolution clause stipulates that arbitration must be conducted in Hong Kong—the cost and difficulty of protecting your rights will increase significantly.
Before renting a server that doesn't require registration, take 30 minutes to carefully read the SLA, focusing on the following 8 aspects:
1. How many nines are in the availability commitment?
2. How is compensation calculated, and what is the maximum?
3. In what situations is compensation not provided?
4. How fast are the response and repair times?
5. Who controls the monitoring data?
6. Will maintenance secretly deduct from your availability?
7. How to recover lost data?
8. Where to resolve disputes?
Jtti always adheres to a transparent and responsible service philosophy. On Jtti's official website product pages and user agreements, you can clearly view the SLA terms for each service—including core content such as availability commitments, compensation mechanisms, and technical support response standards. Jtti is committed to ensuring that every user fully understands the service guarantees they will receive before renting a server that doesn't require registration.
