Servers rarely pick a convenient moment to fail. The outage lands on a Friday evening, in the middle of an email campaign, or during month-end close. That is when "99.9% SLA" and "support 6 days a week" stop being brochure lines and become the only thing that matters. This article puts numbers on those commitments at Open Vista, states plainly what they do not cover, and lists what to check before you sign.
What 99.9% uptime means in real hours
An SLA (service level agreement) of 99.9% does not mean "never down." It means that over a reference period, the infrastructure stays reachable 99.9% of the time. The other 0.1% translates into hours you can count:
- Per year: about 8 hours 45 minutes of allowed downtime.
- Per month (30 days): about 43 minutes.
- Per week: about 10 minutes.
Put differently: if your NVMe VPS is unreachable for half an hour in a year, you are still inside the agreement. If downtime runs for days, the SLA is breached, and the contract should say what happens next: credit, extension, partial refund. That clause is what to read, not the percentage in large type.
One clarification matters. At Open Vista, the 99.9% applies to infrastructure: the host, the network, the storage, the datacenter power, the availability of the virtual machine. It does not apply to what you install on top. A database flattened by an unindexed query, a module that crashes after an update, a site broken by a plugin: those are application incidents, not hosting incidents. The distinction is not legal hair-splitting. It decides who has to act, and how fast.
What support covers, and what it does not
Support answers six days a week. Monitoring runs around the clock. The asymmetry is deliberate and healthy: the system detects, a human acts. Here is what comes with every VPS plan, from VPS-1G at $9.99 a month to VPS-5G at $49.99 a month.
- 24/7 monitoring of the infrastructure, with alerting when something breaks.
- The management panel: web console, reboot, reinstall, DNS management.
- Automatic backups and DDoS protection at no extra cost.
- A firewall, root access, one dedicated IPv4 address and unmetered bandwidth.
- Delivery in under 15 minutes after ordering, and an upgrade path to a bigger plan without a painful migration.
What support will not do for you: build your application, fix a broken theme, write your business rules, tune a slow SQL query, repair a mailbox whose domain was misconfigured, or answer your customers. Those requests are scoped and quoted separately. Saying so upfront prevents the real source of frustration, which is a ticket opened for an application problem that comes back two hours later marked out of scope.
The five VPS plans: what the price actually buys
Price does not buy a better SLA. Across all five plans, monitoring, backups, DDoS protection and the 99.9% commitment are identical. What you pay for is resources, and therefore the ability to absorb load.
| Plan | vCPU | RAM | NVMe | USD/mo | MAD excl. VAT/mo |
|---|---|---|---|---|---|
| VPS-1G Démarrage | 1 | 1 GB | 30 GB | 9.99 | 94.48 |
| VPS-2G Business | 1 | 2 GB | 40 GB | 19.99 | 189.06 |
| VPS-3G Performance | 2 | 3 GB | 50 GB | 34.99 | 330.93 |
| VPS-4G Avancé | 2 | 4 GB | 80 GB | 39.99 | 378.22 |
| VPS-5G Ultra | 3 | 5 GB | 100 GB | 49.99 | 472.80 |
Quick read: VPS-2G ($19.99) doubles the RAM for $10 more, the most cost-effective jump for a brochure site with a database and a cache. VPS-3G ($34.99) adds a second vCPU and 10 GB of NVMe, the plan where an Odoo instance, NextCloud or a small ERP with a handful of concurrent users fits. Above that, VPS-4G ($39.99) and VPS-5G ($49.99) are for workloads that compile, index, or run several containers side by side.
Latency: the datacenter is in New York
Every plan is hosted in New York, USA, the only production location today. A Rabat datacenter has been announced but is not open, so do not plan around it yet. Rough latency figures from New York: under 5 ms to New York, about 8 ms to Ashburn, 10 to 20 ms to Montreal, about 14 ms to Toronto. That is comfortable for ordinary web traffic, including visitors in Morocco. It is not comfortable for voice or trading.
Monitoring, backups and DDoS protection: the invisible layer
What separates decent hosting from sloppy hosting never shows up on a sales page. Three mechanisms do the quiet work.
24/7 monitoring checks that the machine answers, the network holds, and core services are up. An alert fires and someone looks, even when commercial support is observing its six-day week.
Automatic backups are included on every plan. But a backup you have never restored is not a backup, it is a file. The habit is simple: once a quarter, restore a random file into a test directory and confirm it opens.
DDoS protection and the firewall absorb the constant background noise of the internet. They do not replace application security. A weak admin password, an unpatched module, or a carelessly filtered contact form remain your risk, not the network's.
Four things worth testing yourself
- Reboot from the panel in the middle of a workday, and time the actual recovery.
- Restore a backup file, not just create one.
- Reinstall a throwaway OS image, to confirm you can rebuild from scratch.
- Time support's first reply on a Tuesday morning, and again on a Saturday.
What to put in your hosting contract
An SLA is only worth something if it can be measured and enforced. Before signing, get these in writing:
- The exact scope: infrastructure only, or the application too.
- The calculation method: reference period, what counts as downtime, what is excluded (announced maintenance, upstream network failure, force majeure).
- The remedy: credit, extension or refund if the target is missed.
- Support response times, kept separate from the availability SLA, because they are two different promises.
- Backups: frequency, retention, restore procedure, and what happens if you leave.
- Exit: how you get your data and IP addresses back, within what delay and at what cost.
If a provider refuses to pin down those six points, that says less about their engineering than about their contract.
FAQ
Does the 99.9% SLA cover my website?
Not directly. It covers the infrastructure: the virtual machine, the network, the storage. If your site goes down because an update broke the code, the infrastructure was still available and the SLA was not engaged.
How long until I get my VPS?
Delivery is advertised at under 15 minutes after ordering, with immediate root access and a management panel ready to use.
Can I move to a bigger plan without migrating?
Yes. Upgrading is designed to happen without a painful migration: you do not rebuild the server, you give it more resources.
Is support available on Sundays?
Support is staffed six days a week. Infrastructure monitoring runs 24/7, so a hardware or network incident is detected even outside support hours.
Conclusion: measure before you promise
An SLA is not a slogan, it is arithmetic. 99.9% over a year is about 8 hours 45 minutes of allowed downtime on the infrastructure, no more and no less. Everything else depends on what you put on the server, how you back it up, and how clearly the contract is written. To compare configurations, check what is included and pick the right size before ordering, the NVMe VPS page details each plan, and the offer catalogue covers managed hosting, business applications and the services around them. One last useful step: open a test ticket before you pay, and watch the reply. It is often the best predictor of the support you will get six months from now.