SLA and uptime: what hosts promise

An SLA of 99.9% allows a host 43 minutes of downtime a month, 99.99% about four. But a percentage in an advert is not yet a commitment: the commitment is the document that says what counts as downtime, what is excluded and how much you get back. Most often that is a discount on the next invoice, not money. Here is how to read such documents and how to check availability yourself.

Updated 05.11.2026 · Find a VPS for your needs →

An SLA (Service Level Agreement) is the part of the contract in which a host promises a certain level of availability and describes what happens if it fails to deliver. Uptime is the share of time the service was working. Below is how to turn percentages into minutes, what hosts actually promise, what exactly an SLA compensates and why you will not get compensation without your own monitoring.

What the percentages mean in minutes

The difference between 99.9% and 99.99% looks like rounding, but in minutes it is tenfold. For a 30-day month and a year, not counting planned maintenance:

Availability	Allowed downtime per month	Per year
99%	7 h 12 min	87 h 36 min
99.5%	3 h 36 min	43 h 48 min
99.9%	43 min 12 s	8 h 46 min
99.95%	21 min 36 s	4 h 23 min
99.99%	4 min 19 s	52 min 34 s
99.999%	26 s	5 min 15 s

A single three-hour outage in a 30-day month brings availability down to 99.58%. Rebooting a host to update its kernel takes several minutes, so 99.99% for a single physical server is an almost impossible promise: it becomes realistic when machines run on several hosts and are moved between them without stopping. Mind the length of the month too: in a 31-day month, 99.9% already means 44 minutes 38 seconds.

What hosts promise

According to the accsly.io catalog as of November 2026, 1,479 of 2,242 VPS hosts state their availability, and 1,409 of them give a specific percentage. Taking the highest figure each host names, the picture is as follows:

99.9%: 720 hosts, or 51%;
above 99.9% but below 99.99% (usually 99.95% or 99.98%): 161, or 11%;
from 99.99% to 99.999%: 327, or 23%;
100%: 86, or 6%;
below 99.9%: 115, or 8%.

At 150 hosts the description contains two or more different figures: one on the home page and another on the plan page or in the terms, sometimes a different one for the network and for the servers. Only 62 hosts mention compensation in the terms recorded in the catalog. That does not mean the others have none, but it does mean you will not find it on the visible pages of their site: you need the document itself.

What an SLA actually pays

An SLA almost never covers your losses; it returns part of the fee for the service, and most often not as money but as a discount. This is easy to see in the documents of large cloud platforms that publish their SLAs in full.

The SLA of one international cloud platform promises 99.9% for a single virtual machine and, if this is breached, gives a credit of 10% of the monthly bill for that machine when availability was between 95% and 99.9%, 25% between 90% and 95%, and 100% below 90%. The credit goes towards future payments and cannot exceed the monthly bill for the service.
The SLA of the largest international platform promises only 99.5% for a single machine, and 99.99% for machines deployed across several availability zones. The credits are 10%, 30% and 100% on similar steps, and they are not paid out as money.
A large Russian cloud's agreement states plainly that compensation is a discount that reduces the cost of services, not a cash payment, and that it cannot exceed the cost of the service for the billing period.

Let us work through an example. A VPS at $21 a month was down for three hours, so availability was 99.58%. Under the first scheme you are entitled to 10%, that is a $2.10 discount on next month. Under the second, with a 99.5% threshold, you get nothing: three hours fit within the promise. If an online shop lost orders during those three hours, the SLA will not cover them. So an SLA is a lever on the host and a signal of how confident it is in its infrastructure, not business insurance.

Exclusions and claim deadlines

The key part of an SLA is not the percentage but the list of what does not count as downtime. Typical items appear in almost every document:

Planned maintenance. In the Russian cloud's agreement from the example above, maintenance announced three days in advance is not counted as an interruption, and downtime caused by preventive and emergency work of up to 8 hours per billing period is not compensated.
Customer actions. A misconfigured firewall, a full disk or a crashed application is not the host's downtime.
External causes. Force majeure, orders from government authorities and often DDoS attacks. What changes during an attack is covered in the article DDoS protection at hosts.
Rule violations and unpaid bills. If the server was suspended for breaking the terms of use or the invoice is overdue, there will be no compensation.
What exactly is measured. Some SLAs measure the availability of the data centre network, others that of the host, others that of your machine. Under the international cloud's document, downtime means losing external connectivity or access to the disk, and only periods of one minute or longer count.

Compensation is almost never applied automatically: you have to claim it. The deadline is limited: 30 calendar days from the incident in the Russian cloud's agreement, 60 days at one international platform, and the end of the second billing cycle at the other. The claim must include the date and time of each incident, and the international documents explicitly require logs that show the downtime. Miss the deadline and you lose the right to compensation.

Uptime without an SLA document

"99.99% uptime" on the home page and an SLA are different things. The figure in an advert promises nothing unless there is a section of the contract or a separate document behind it. Before buying, find that document and check:

over what period availability is measured: a month, a quarter or a year;
what counts as downtime and from what minimum duration;
which work and events are excluded and how planned maintenance is announced;
how much is compensated at each availability step and whether there is a cap;
in what form: money, account credit or extra days;
how and by when to file a claim and what evidence is needed.

If there is no such document, ask support in writing. An answer like "everything is stable here" without a reference to the terms means there is no availability commitment. That does not necessarily make it a bad host, but comparing it by percentage is pointless. A promise of 100% is also worth a question: what will the host give back when that percentage is inevitably missed?

How to check availability yourself

The host measures availability with its own data, and the customer's data often settles a dispute. You need an external check that does not run on the same server, and preferably not at the same host: an HTTP or ping probe once a minute, a log with the exact start and end time of every failure, and an alert to a messenger or email. How to set this up for free is described in the article Server monitoring: uptime, load, alerts, and plans for a monitoring server are collected in the selection VPS for monitoring.

If an outage happens, save the log and file the claim straight away rather than at the end of the month: state the start and end time, the server's address and what exactly was unavailable. A month with several short outages is also worth adding up, as the total may cross the SLA threshold.

If availability really matters

No SLA makes a single server more reliable: it only distributes responsibility. If an hour of downtime costs more than the monthly fee for the server, build failure into the architecture:

keep a backup on another server and test restoring it, see the article VPS backups;
for an important service, run a second server in another location or at another host and switch traffic via DNS or a load balancer;
on a physical machine, put the disks in RAID and find out how quickly hardware is replaced; the differences from a VPS are covered in the article Dedicated server or VPS.

Two 99.9% servers at different hosts, with independent failures, together give more than one server promised at 99.99%, and they usually cost less than one server with real guarantees at that level.

FAQ

How much downtime does a 99.9% SLA allow?

43 minutes 12 seconds in a 30-day month and about 8 hours 46 minutes a year. For 99.95% it is 21 minutes 36 seconds a month, and for 99.99% 4 minutes 19 seconds.

What does a host pay if it breaches the SLA?

Usually a share of the monthly fee for the affected service, for example 10%, 25% or 100% depending on the actual availability. Most often it is a discount on the next invoice rather than money, and the amount does not exceed the monthly fee.

Does an SLA compensate business losses from downtime?

No. An SLA returns part of the fee for the service; lost revenue and lost data are not covered. Backups and a second server protect you from losses, not an agreement.

What uptime do VPS hosts promise?

According to the accsly.io catalog as of November 2026, 51% of the hosts that give a percentage promise 99.9%; 29% promise 99.99% or more, and 6% claim 100%. Only a handful mention compensation on the visible pages of their site.

How do I get compensation under an SLA?

Write to the host within the deadline set in the agreement, usually 30 to 60 days, with the date and time of each outage and the log from your external monitoring. Compensation is rarely applied automatically: without a claim you usually get nothing.

Does planned maintenance count as downtime?

As a rule, no, if the host announced it in advance in the way the SLA describes. Some agreements also do not compensate emergency work up to a certain total duration, for example up to 8 hours per billing period.

Plans on this topic

VPS for Asterisk and VoIP VPS for blockchain nodes VPS for an Ethereum node Best VPS hosting in Russia 2026

More in the guide

CDN: when a website needs oneA CDN stores copies of a site's files on servers around the world and delivers them to each visitor from the nearest one. This speeds up loading of images, scripts and video and takes load off your server, but it does not make a slow CMS faster and does not replace choosing the right country for your server. Here is when a CDN is really needed and when a single VPS in the right place is enough.
Domains and DNS for a server ownerYou can switch servers in an evening, but a domain stays with a project for years. Through DNS, the domain points to your server, your mail and third-party services, and a single wrong record can take the site offline or send your emails to spam. Here are the records a VPS owner will need, how caching works, where to keep the zone and how to move to a new server without visitors noticing.
SSL certificates: free and paidA free certificate encrypts the connection just as reliably as a paid one: browsers make no distinction between them in terms of security. The difference lies in what the certificate authority checks, how long the certificate lasts and who is responsible for renewing it. According to the accsly.io catalog as of November 2026, free SSL is explicitly listed in 1,449 shared hosting plans from 178 hosts. Here is when that is enough and when a certificate is worth paying for.
152-FZ and hosting: where to store the dataIf a website or service collects personal data of Russian citizens, that data must be recorded and stored in databases located in Russia: this is Part 5 of Article 18 of Federal Law No. 152-FZ. For choosing a VPS, this means the data centre's country matters, not the host's, and copies abroad fall under separate rules. Here is what the law requires, what duties fall on the server owner, how it applies to companies outside Russia and which setups are used in practice.
GDPR and choosing a hostThe EU General Data Protection Regulation (GDPR) does not require data to be kept in the EU, but it does require a data processing agreement with the host and a lawful mechanism for any transfer of data out of the EU. According to the accsly.io catalog as of November 2026, 61% of VPS hosts have servers in the EU, but almost half of them are registered outside the European Economic Area. Here is why that matters and what to check in the terms.
Hosting provider jurisdiction: what it affectsEvery server has two countries: the one where the data centre stands and the one where the company renting it out is registered. According to the accsly.io catalog as of November 2026, these are different countries for 15% of VPS hosts. The first determines latency and the rules on where data is stored; the second determines the law of the contract, the court for disputes, the documents your accountant receives and whose authorities can request your data. Here is what each of them changes.
How to compare VPS prices: the price of a resource unitA plan's price tells you how much you will pay, but not how much you get for that money. According to the accsly.io catalog as of November 2026, a gigabyte of memory on a 1 GB plan costs a median $8.00 a month, while on a 4 GB plan it costs $5.00. Here is how to work out the price per vCPU, per gigabyte of memory and of disk, and what to bring to a common basis before you compare two offers.
Free trials and money-back guaranteesThere are two ways to test a server before paying for a month or a year: a free trial or a money-back guarantee. According to the accsly.io catalog as of November 2026, 11% of VPS hosts offer a trial and 36% a money-back guarantee. But at many of them the terms apply only to shared hosting, only to the first order or only to card payments. Here is what to check so that the guarantee actually works.
Cloud server or VPSA cloud server and a VPS are usually the same KVM virtual machine. The difference lies in what surrounds it: a cloud sells resources by the hour through an API and offers network disks, snapshots and ready-made services, while a classic VPS is a fixed monthly plan with a local disk and a traffic allowance. Here is which of this you actually need and what it costs.
Dedicated server or VPSA VPS is a slice of a physical server, a dedicated server is the whole machine. With 4 cores and 8–16 GB of memory a VPS costs almost half as much, with 4 cores and 32 GB the prices are equal, and with 64 GB on 8 cores or 128 GB on 16 cores a dedicated server in the catalog is a third cheaper than a virtual one. Here is what else changes besides the price: neighbours, guarantees, delivery time, hardware failures and remote management.
Shared hosting or VPS for a websiteShared hosting is a ready place for a site where the host looks after the server. A VPS is your own server where you set everything up. For a landing page or an ordinary WordPress site, hosting is enough and costs several times less. You need a VPS when the account limits are too tight for the site or when it needs software that hosting does not allow. Below is what each option includes, what it costs and a table to help you choose.
DDoS protection: what hosting providers actually give youEvery second host now lists "DDoS protection" in its plans. Behind that line there may be multi-terabit filtering, or only the right to take your address offline when an attack bothers the neighbours. Here is what protection is made of, how L3/L4 differs from L7, and how to tell from the contract what you are buying.
Mail server on a VPS: SPF, DKIM, DMARCInstalling Postfix or a ready-made bundle like Mailcow on a VPS takes an hour. Getting your mail into the inbox rather than spam is harder: mailbox providers look at the port, the server's address, its reputation and three DNS records. Here is what to check with the host before buying and how to set up SPF, DKIM and DMARC.
Server monitoring: uptime, load, alertsWithout monitoring you learn that the server is down from your visitors, and that the disk is running out when the database stops writing. A few checks from outside and a dozen metrics from inside are enough to see a problem before users do. Here is what to measure, where to run checks and how to do it for free.
How to read hosting plans: the hidden limitsA plan card shows cores, memory, disk and a monthly price. Yet two plans with identical numbers can cost different amounts a year later and behave differently under load. The difference lies in terms printed in small type, in the acceptable use policy, or only in the contract. Below are eight places worth checking before you pay.
IPv4 and IPv6 on a serverA typical VPS has one public IPv4 address and, if the host supports it, an IPv6 subnet. An extra IPv4 address almost always costs money, and the cheapest plans sometimes come without their own IPv4 at all. Here is where the shortage comes from, what you can do with IPv6 alone and which address settings a mail server needs.
Bandwidth and traffic: what 1 Gbps and unlimited really meanA server plan usually has two network lines: port speed and traffic volume. The first says how fast the server can send data, the second how much it can send in a month. They are linked by simple arithmetic, and once you know it, it is easy to see what "1 Gbps unlimited" actually offers and whether "5 TB" will be enough for you.
Docker on a VPS: getting startedDocker lets you run a bot, a website, a database and monitoring on one server without them getting in each other's way, and moving to another VPS comes down to copying a couple of files and the data. Installation takes five minutes; the real trouble starts later: database ports exposed to the internet, logs growing to tens of gigabytes and data that vanished along with a container. Here is how to avoid all of that from the start.
How to move a website from shared hosting to a VPSThe move itself takes a couple of hours; the mistakes usually happen around it: a forgotten cron job, .htaccess rules that nginx does not read, an AAAA record with the old address, or mail that disappeared along with the IP change. Below is the order of steps with commands and checks for a PHP and MySQL site.
Server control panels: ISPmanager, FASTPANEL, aaPanelA panel turns a VPS into the kind of dashboard you know from shared hosting: sites, databases, mail and certificates are set up with a few clicks instead of editing config files. The three panels most often installed on Russian and budget international servers are ISPmanager, FASTPANEL and aaPanel. They differ in price, memory requirements and how much of the server they take over.
Windows VPS for 1CA rented server for 1C saves you from the office PC under a desk that nobody updates or backs up. But 1C on a VPS has its own quirks: the software licence is tied to the parameters of the virtual machine, users need both 1C and Windows licences, and core speed matters more than core count. Here is how to put together a 1C server for a small company.
Windows VPS: licensing, RDP and secure sign-inA Windows server costs more than Linux not only because of memory: the price includes a licence, and how it is arranged decides whether you are compliant and how many people can connect. An exposed RDP port is also the most common way such servers get hacked. Here are both topics.
When to move from shared hosting to a VPSA VPS does not make a site faster by itself: it gives you resources and freedom, and with them the job of running a server. Here is how to tell that hosting has become too small, when the problem can be solved without moving, and how to migrate a site so visitors notice nothing.
VPS backups: strategies and toolsOn a VPS, backups are your job, even if the plan says "backups included". We go through what exactly to back up, why a host snapshot is not enough, how to set up restic or BorgBackup in half an hour, and how to make sure you can actually restore from a backup.
First 10 steps after buying a VPS: securityBots start guessing passwords on a new server within minutes of it getting a public IP. Below are ten steps that close the most common holes in about an hour: commands for Ubuntu and Debian, the differences for AlmaLinux and Rocky Linux, and the checks that keep you from locking yourself out of your own server.
How to choose a country for your serverPut the server where your visitors are. For an audience across Europe, Germany, the Netherlands and Finland are convenient; for North America, the US East Coast. If you serve EU users, keeping the data in the EU makes GDPR simpler. Price hardly decides it: at the same host a server in Frankfurt, Amsterdam or New York usually costs the same, and only Asia and Australia are noticeably more expensive.
Hourly billing: when it pays offHourly billing pays off if you need the server for less than three weeks a month: for a test, a build, a one-off heavy job. At most hosts an hour costs exactly the monthly price divided by 730 hours, so over a full month you hardly overpay. The overpayment hides elsewhere: hourly plans are dearer on average, and for a server that runs all year, monthly billing or a year paid in advance usually comes out cheaper.
How much RAM and how many cores a website needsMost websites are fine with 2 GB of memory and one or two cores; a shop with thousands of products needs 4 GB and two cores. Memory matters more than cores: when it runs out, the site slows down or crashes, while an extra core without memory adds almost nothing. Below I show how to work out the memory for your site and how to see what it actually uses.
KVM or OpenVZ: which to chooseIn 2026 KVM is almost always the right pick: it costs a dollar or two more per month than OpenVZ, but everything runs on it, from Docker to Windows. OpenVZ makes sense in one narrow case, explained below, and at the end you will find the commands to check your own server in a minute.
NVMe or SSD: does it matter for a websiteFor a small site the difference between NVMe and SSD is barely noticeable, because pages are served from memory rather than from disk. It shows up in shops and sites with a large database. At the same host NVMe costs only about $0.50–1 a month more, and sometimes less, so what usually decides it is not the disk type but its size and how the host has limited it.
What is a VPS, in plain EnglishA VPS is a slice of a powerful physical server that you rent and that behaves like a separate computer, with its own operating system, its own IP address and full control. Below is how it differs from shared hosting, who needs one, what it costs and what to check right after you buy it.
VPS for a Telegram bot: what the bot needs and how to pick a planA bot with a menu and a few commands runs fine on the cheapest server, while a bot with broadcasts, files and a database hits completely different limits. Here is what a bot really consumes, which Telegram limits shape the choice of server and how to avoid overpaying.
How to test VPS ping and speed before you payThe numbers in a plan do not tell you the latency to your users or how fast the disk really is. Most of this can be measured before you pay, the rest in the first hour after launch. Here are the tools and how to read the results.
Corporate network on a VPS: remote access to the officeA rented server can become the hub of a company's private network: remote employees and branch offices connect through it to the ERP, file server, CRM and Git. Here are the common layouts, protocols, server requirements and security basics.
How to choose a VPS: the 7 specs that matterTwo plans with the same "2 cores, 4 GB, 50 GB" can differ several times in speed and twice in price a year later. Here is what stands behind each line of a VPS plan and what to check before you pay.
VPS vs VDS: is there a difference and what actually mattersIn most markets VPS and VDS mean the same thing: a virtual server. The difference that really affects your work is hidden not in the name but in the virtualization type and whether resources are guaranteed.