Skip to content
Find my data room
Costs and contracts

Data room uptime and SLAs: what 99.9% means the night before a bid deadline

A 99.9% uptime promise still allows about 43 minutes of downtime a month. Here is the math for each level, what SLAs usually exclude, and why service credits rarely cover the real cost.

Uptime figures look reassuringly similar across providers. Almost everyone claims something close to 99.9%. The differences are in the definitions, the exclusions, and what happens when the target is missed, and they matter most at the worst possible moment: the evening before final bids are due.

How much downtime each uptime percentage allows in a month

99% 7.2 h
99.5% 3.6 h
99.9% 43 min
99.95% 22 min
99.99% 4.3 min

The common data room promise, 99.9%, still allows about 43 minutes of unplanned downtime every month. That is one bid-deadline evening, if it lands at the wrong time.

The percentage matters less than three details: what counts as downtime, what is excluded, and what you get back when the target is missed.
dataroomreview.org
Calculated for a 30-day month (43,200 minutes). Most SLAs exclude scheduled maintenance from the count, so read the definition before comparing numbers.

What does a data room uptime SLA actually promise?

A service level agreement sets a measurable availability target over a period, usually a calendar month, and a remedy if the provider falls short. It is a contract term, so the precise wording counts. Three definitions sit underneath every percentage:

  • What “available” means. Can users log in? Can they open documents? Some SLAs count only the login page.
  • How it is measured. Provider-side monitoring, external probes, or customer reports.
  • What is excluded. Scheduled maintenance, customer-side problems, third-party outages, and force majeure are common carve-outs.

A 99.9% SLA with broad exclusions can promise less than a 99.5% SLA with narrow ones.

How much downtime does each uptime level allow?

The math is simple. A 30-day month has 43,200 minutes.

Uptime targetDowntime per 30-day monthDowntime per year
99%7.2 hoursabout 3.65 days
99.5%3.6 hoursabout 43.8 hours
99.9%43.2 minutesabout 8.8 hours
99.95%21.6 minutesabout 4.4 hours
99.99%4.3 minutesabout 53 minutes

These are ceilings, not predictions. A well-run platform may go months without an incident. The SLA tells you what the provider is prepared to be held to, which is a different thing from what it usually achieves.

43,200minutes in a 30-day month
43 minallowed at 99.9% uptime
4 minallowed at 99.99% uptime

Arithmetic on the SLA percentage. Scheduled maintenance is usually excluded and comes on top.

Why does scheduled maintenance matter so much?

Because it usually does not count. If a provider reserves a weekly maintenance window, that time sits outside the SLA entirely. For a domestic deal this is rarely a problem, since windows are typically placed overnight in the provider’s main region. For a cross-border auction with bidders in Asia, Europe and North America, someone’s working day overlaps with almost any window.

Ask three questions: when are maintenance windows, how much notice is given, and can maintenance be deferred around a known deadline? Providers serving large transactions often agree to freeze changes during critical dates if asked in advance.

What are service credits worth?

A service credit is the standard remedy: a percentage discount on a future invoice, scaled to how badly the target was missed. Credits typically have to be claimed within a set period and are often capped at a fraction of the monthly fee.

The limitation is obvious once stated. If a room goes down for two hours the night before final bids, a credit of a few percent of one month’s fee is not compensation for the deal risk. Credits are a signal that the provider takes availability seriously, not insurance against a failed process. Treat the SLA as one input alongside the provider’s architecture and track record.

Plan for an outage you cannot prevent

For any hard deadline, agree a fallback in the process letter: how bids or final documents can be submitted if the room is unreachable, and who confirms receipt. It costs nothing to write down and removes the panic if it is ever needed.

What else should you ask about availability?

QuestionA reassuring answerA weak answer
Is there a public status page?Yes, with incident history”Contact support if there is an issue”
Where is the platform hosted?Named regions, with a choice where neededNo detail
What are the recovery targets?Stated recovery time and recovery point objectives”We have backups”
How are incidents communicated?Proactive email and status updatesOnly on request
Is there a change freeze option?Yes, for agreed critical datesNot offered
Is support available at the deadline?Around the clock, with escalationBusiness hours in one time zone

Recovery targets deserve a word. A recovery time objective is how quickly service should be restored after a serious failure. A recovery point objective is how much recent data could be lost. For a data room, losing the last hour of uploads and Q&A answers before a deadline would be painful even if the platform came back quickly.

How do support and uptime work together?

An outage is often a partial one: a slow viewer, a failing upload, a region with login problems. What matters then is whether someone answers quickly and knows what to do. Our guide to data room customer support covers channels, coverage and what “24/7” usually means in practice, and our review method weighs service quality as one of six scoring criteria.

Where do you find a provider’s SLA?

Self-serve providers often publish their SLA or reference it in their terms of service. Enterprise providers usually attach it to the master agreement or order form, sometimes with different tiers for different contract sizes. If you cannot find one, ask for it in writing before signing; a provider that will not commit to any target is telling you something.

Our reviews cover each provider’s service and security posture, and our pricing guide lists contract lines that sit alongside the SLA, such as extension rates and archive fees.

Is 99.99% worth paying more for?

For most transactions, the difference between 99.9% and 99.99% on paper is less important than maintenance exclusions, deadline support and a written fallback plan. A higher number with broad exclusions protects you less than a slightly lower one that counts every minute users cannot work. Compare definitions first, percentages second.

Frequently asked questions

What does 99.9% uptime mean for a data room?

It allows about 43 minutes of downtime in a 30-day month, or roughly 8.8 hours a year, usually excluding scheduled maintenance. It is a contractual ceiling, not a forecast.

Does scheduled maintenance count against the SLA?

Usually not. Most SLAs exclude announced maintenance windows, so ask when they occur, how much notice is given and whether they can be moved around your deadlines.

What is a service credit?

A discount on a future invoice when the provider misses its uptime target. Credits are usually capped and must be claimed within a set time, so they rarely cover the business impact of an outage.

How can a deal team protect itself against data room downtime?

Check the SLA definitions and maintenance windows, ask for a change freeze around key dates, confirm deadline support coverage, and write a fallback submission method into the process letter.

What are RTO and RPO?

The recovery time objective is how quickly service should be restored after a serious failure; the recovery point objective is how much recent data could be lost. Both should be stated by an enterprise-grade provider.

More from the blog