Uptime SLAs explained
18 Sep 2023
Meet Derick, in-house legal counsel for Vector AI, a rapidly growing tech company.
Derick is your one-person legal team – always with a coffee mug in hand, a joke on his lips, and sporting a curious collection of action figures on his desk. You know the type.
So, one day, Derick is called into the Boss’s office.
The Boss explains that Vector AI needs to enter into a mission-critical cloud services agreement with a vendor, and the contract must adequately address uptime.
Derick’s Version of “uptime”:
Derick thought he had it in the bag. “Uptime? Oh, you mean like my personal record of staying up for 36 hours straight during the ‘Call of Duty’ launch weekend? I got this!” Ah, if only.
Eager to get going, Derick grabs his old university textbooks and notes on contract drafting.
Nothing. Zilch. Absolutely no mention of “uptime”.
No worries. Google to the rescue. As it turns out, uptime means something completely different than what Derick had in mind.
As he read through terms like “SLAs,” “availability,” and “99.9% uptime,” he quickly realised he might be in over his head. What he thought was a simple equation of wakefulness versus sleep was more complicated than explaining why cats hate water.
Jen to the rescue.
Derick phones up Jen. She was the smartest girl in class.
Jen tells Derick, “Imagine if uptime was like your favourite pizza delivery place promising to deliver your pizza 99.9% of the time within 30 minutes or less. Missing that mark means you get free garlic knots on your next order. Eureka!” The analogy was far from perfect, but it was a start.
Jen then tells Derick there is a secret way to draft these contracts without spending countless hours researching, copying and pasting from old contracts. He just has to Ninja it!
Jen explains that ContractNinja is an advanced contract automation system that allows you to build complex tech-related contracts in a guided and intuitive manner. And if you get stuck, ContractNinja has a free online resource that explains these contracts in an easy-to-understand way.
Derick’s quest to master tech contracts starts.
As it turns out, “uptime service levels” are not as daunting as initially thought! You just need to break it down.
Let’s dive in!
What is uptime?
When we talk about “uptime,” we’re talking about how often something is working as it should be.
With cloud service, you want it running smoothly so you can, say, access your files, run your website, or conduct whatever business you need to do on the cloud without throwing your computer out the window in frustration.
Uptime is usually expressed in percentages; trust me, that decimal point can make a world of difference. For example, a 99% uptime sounds good on paper. But when you translate that into real-world time, it means the service could be down for about 5256 minutes a year without any consequences for the vendor. Yikes!
On the other hand, a 99.999% uptime – nicknamed “five nines” – means the service is allowed to hiccup for just about 5 minutes per year. Suddenly, percentages got a whole lot more interesting, huh?
Importantly, when considering uptime, you must be precise about calculating it.
Certain cloud providers exclude downtime that is less than a certain period. So, if, for example, the system is down for less than five minutes, it will not be regarded as downtime.
Then, there is scheduled maintenance. These scheduled maintenance periods generally don’t count towards downtime. But, as with everything, it has to be within reason.
From the customer perspective, you want to provide that notice must be provided before any scheduled maintenance, and if the cloud service is critical to your business, you also want to provide that scheduled maintenance cannot be performed during certain operating times of your business.
Another important aspect to consider is when will the downtime calculation start. Ideally, as a cloud service provider, you want the calculation to start once a ticket has been logged.
What happens if the service provider fails to meet the uptime service levels? Generally, the service provider will give credits to the customer.
Credits basically give the customer a discount on their next invoice.
There are, however, a couple of finer details to consider when it comes to credits.
- How will the credits be calculated?
- Are there any limits placed on the credits? For example, credits will only be provided to a maximum of 10% of the next month’s invoice?
- Will the credits be the sole remedy available to the customer?
- To receive credits, must a claim for credits first be logged with the service provider?
- If a request for credits must first be logged, must the request for credits be logged within a certain period to be valid?
- Will credits be forfeited on termination?
- With credits, the devil is in the detail. So be sure to review the calculations carefully!
Let’s consider Vector AI, our example company, a business that will rely heavily on the cloud services.
They sign a contract with the provider who enticed the conclusion of the contract with the promise of 99.9999% uptime. Everything seems perfect on paper. However, in the first quarter, the provider suffers two significant outages, each lasting a couple of hours, rendering Vector AI unable to provide services to its customers.
Despite apologies and SLA credits, the damage for Vector AI is significant. Not only have they lost productivity, but they’ve also damaged their reputation with customers. At this point, they’re not looking for more SLA credits; they’re looking for a way out.
Enter the termination right.
Suppose you want to include a termination right; ensure you are precise and provide what will trigger such a termination right – for example, failure to meet the uptime service levels for three consecutive months.
The above is a lot to consider! Luckily, Derick found ContractNinja!
- Understanding tech buzzwords like AI
- The indispensable tool of legal technology
- How AI-powered legal assistance is changing the game
- Unpacking tech lingo and why you cannot ignore it