Cloudways is a managed cloud hosting platform that sits on top of five infrastructure providers: DigitalOcean, Vultr, Linode, AWS, and Google Cloud. Rather than managing a server directly, you work through Cloudways’ dashboard, which handles the stack, security patching, and automated maintenance.

To test it properly, I signed up for a real account, provisioned a 1GB WordPress server on DigitalOcean’s Premium NVMe tier in New York, and ran it for 30 days.

Performance monitoring ran daily through GTmetrix and continuously through UptimeRobot. I also tested live chat support with a technical question that required more than a documentation link to answer.

The result is a platform with real strengths in uptime and ease of use, and a few limitations worth knowing before signing up.

Best for Growing Businesses, Web Developers, and Agencies

Unparalleled speed, robust reliability, and tailored cloud hosting packages from Cloudways are built with modern solutions in mind. Ideal for all web hosting needs, and extremely user-friendly.

Visit Cloudways

Cloudways Pros and Cons

Cloudways Rating Breakdown

The scores below reflect my hands-on findings across every area I tested during this review: provisioning, daily performance monitoring, support quality, and the overall dashboard experience.

Each parameter is scored out of 10 based on real results rather than marketing claims. For a full explanation of how HostAdvice scores hosting providers, see the HostAdvice Rating Methodology.

Best for Growing Businesses, Web Developers, and Agencies

Unparalleled speed, robust reliability, and tailored cloud hosting packages from Cloudways are built with modern solutions in mind. Ideal for all web hosting needs, and extremely user-friendly.

Visit Cloudways

Cloudways Prices & Plans 2026

Cloudways bills hourly with monthly settlement, so you only pay for what you use. A 3-day free trial runs without requiring a credit card. There is no traditional money-back guarantee: the DigitalOcean Terms of Service state that paid fees are non-refundable.

Payment is by credit or debit card. Pricing varies by cloud provider, with DigitalOcean and Vultr at the lower end and AWS and Google Cloud significantly higher.

Cloudways Features

Cloudways Performance

I tested the DigitalOcean flagship server (HA-Cloudways-DO, 1GB Basic Premium, New York) across four layers:

• Point-in-time page load testing from two geographic locations,
• A 30-day daily GTmetrix monitoring run,
• 30-day uptime monitoring via UptimeRobot, and
• A global HTTP reachability check via Check-Host.

The aim was to answer the question a prospective customer actually cares about: not whether Cloudways looks good on a single test on a good day, but whether the platform performs consistently over real time.

Test server: HA-Cloudways-DO | DigitalOcean Premium (NVMe) | 1GB Basic | New York, USA | WordPress + Lightning Stack (Nginx) + Redis

GTmetrix Point-in-Time: United States

I chose GTmetrix as the primary measurement tool because it tests real-world page load as a user with a browser experiences it, not just a raw server ping. It measures the metrics that search engines and real visitors actually care about, including Largest Contentful Paint, Time to First Byte, Cumulative Layout Shift, and Total Blocking Time.

The server is hosted in New York, so I ran the US test first to establish what the platform delivers at its best, with minimal physical distance between the test node and the server.

Test location: San Antonio, TX | Date: July 17, 2026

• A 92% Performance score on a 1GB shared Basic server is a strong result.
• The TTFB of 179ms reflects the Nginx and Redis stack doing real work: the server is processing requests quickly and returning cached content before the browser has to do much of anything.
• The LCP sitting just under 1 second puts this page in Google’s “Good” threshold.
• CLS of zero means nothing shifts on screen during load, which matters both for Core Web Vitals and for user experience.
• The 2.2-second fully loaded time is the figure to watch, but for a default WordPress install with no CDN on a 1GB server, it is within a reasonable range.

GTmetrix Point-in-Time: Europe

The US result tells one part of the story. Cloudways is used globally, and many customers are based in Europe.

Since the test server sits in New York, European visitors face additional round-trip latency that the US test can’t reveal. I ran a second GTmetrix test from Frankfurt to put a real number on that gap.

Test location: Frankfurt, Germany | Date: July 18, 2026

• The 88% Performance score from Frankfurt still sits in solid territory
• An LCP of 1.3 seconds keeps the page within Google’s acceptable range.
• The TTFB doubling to 356ms is the expected cost of the transatlantic distance: this is physics, not a server problem.

NOTE: European visitors who need sub-200ms TTFB would need a Cloudways server deployed in a European region, which the platform supports.

What this test confirms is that the server itself is not adding unnecessary overhead on top of the network latency, the backend processing time in Frankfurt (171ms) is comparable to what the US server shows (79ms backend), scaled for distance.

30-Day Daily Monitoring

A single GTmetrix test can catch a server on a good day or a bad one. I set up daily automated monitoring from GTmetrix’s San Antonio node running continuously from June 18 through July 17, 2026, 30 consecutive days of data on the same server, same location, same time each morning.

Here are the 30-day averages:

What the data shows:

On 26 of 30 days, the server performed exactly as the point-in-time test suggested:

• Performance scores between 82% and 92%
• LCP comfortably under 2 seconds
• TTFB never exceeding 267ms across the entire month.

The CLS score was zero every single day without exception, a detail worth calling out plainly. Layout stability is one of those metrics that degrades quietly on poorly managed WordPress stacks, and this server never moved a pixel on screen during load across 30 days.

Four days produced a different picture: June 22, June 25, June 28, and July 12. On each of those days, LCP spiked to approximately 3.1 to 3.2 seconds, and the Performance score dropped to 75 or 76%.

The TTFB on those same days stayed normal, between 186ms and 216ms, which rules out a server availability problem or network issue. The backend was responding normally; the delay was in the page rendering itself.

The first three spikes, June 22, 25, and 28, fell exactly three days apart, which suggests a scheduled process running on a 72-hour cycle: likely a backup job, database optimization, or platform-level maintenance task generating enough background load to push LCP over the 2.5-second threshold during that test window.

The July 12 spike breaks the three-day pattern and arrived 14 days after the previous one, so the two clusters may have different causes. In both cases, the server returned to normal the following day with no residual effect.

These four days are worth reporting honestly. A user checking GTmetrix on one of those mornings would see a 75% score and potentially a concern.

UptimeRobot: 30-Day Uptime Monitoring

GTmetrix tells me how fast the page loads. It doesn’t tell me whether the server was actually up and responding every time a real visitor tried to reach it.

For that, I set up a separate UptimeRobot HTTP monitor on the same day the server launched, June 18, checking the WordPress URL every five minutes for the full 30-day period.

The results as of July 18:

100% uptime over 30 days with zero incidents recorded. The server was checked roughly 8,640 times across the monitoring period and returned a successful response every single time.

The response times here, averaging 53ms, reflect UptimeRobot’s lightweight HTTP check rather than a full page load. This is the server confirming it is alive and responding, not GTmetrix measuring how long a browser takes to render the page. The two numbers measure different things and shouldn’t be compared directly.

What the 53ms average does confirm is that the server was not hanging or slow to acknowledge requests at any point during the month, which is consistent with what the GTmetrix TTFB data showed.

Global Reach: Check-Host

GTmetrix tests from one location at a time. To get a snapshot of how the server responds globally, I ran a Check-Host HTTP test against the live WordPress URL, hitting the server from 57 probe locations across five continents simultaneously.

Key findings:

56 of 57 locations returned HTTP 200 OK.

TAKEAWAY: The one connection timeout in Moscow and the 7.94 second response from Iran, Khonj are routing-level issues rather than server failures. Every other location connecting through the same global internet returned a clean response, and Iran’s other cities (Tehran, Qom, Shiraz) all came back under 720ms.

The Khonj result reflects a known routing anomaly on that specific probe path rather than anything the server did.

For a WordPress site hosted in New York with no CDN configured, the global response picture is what it should be: fast in the US, acceptable across Europe, slower but functional in Asia and the Middle East.

Overall Verdict: Cloudways Performance

Best for Growing Businesses, Begineers, and Developers

Unparalleled speed, robust reliability, and tailored cloud hosting packages from Cloudways are built with modern solutions in mind. Ideal for all web hosting needs, and extremely user-friendly.

Visit Cloudways

Cloudways Ease of Use

I worked through the full onboarding flow myself, from a blank browser tab to an active WordPress server on a real paid account. The four subsections below cover what that experience looks like in practice.

1. Registration

The homepage opens with the Lightning Stack pitch, “65% faster performance” in large type, before anything else. I clicked ‘Start Free Trial’ and landed on the signup form.

What the form asks for:

• First name and last name
• Email address
• Password
• What best describes you (dropdown)
• Monthly hosting spend (dropdown)

Those last two are profiling questions, not verification steps. None of the options gate what you can do once you’re inside. Social login shortcuts, DigitalOcean, LinkedIn, GitHub, Google, sit above the form if you’d rather skip the fields entirely.

A Cloudflare check ran quietly in the background and cleared itself without requiring any interaction. The form validated inline rather than on submit, catching my missing first name immediately. One click on ‘Sign Up’, and the account was active.

From opening the signup page to a logged-in account: under two minutes, no card required.

2. Dashboard

Logging in lands you on a personal greeting and a two-card summary at the top:

Below that sits a rotating announcements panel, currently advertising free migrations, and then an Applications list showing every deployed app with its server name, project, and creation date visible without an extra click.

Four Quick Start Guide cards line the bottom of the page covering the questions a new user actually has: how to launch a server, how to migrate, how to add SSL, how to go live.

The sidebar has five top-level links only. Nothing extra competes for attention at first glance, which is a deliberate choice on a platform that otherwise has a lot of moving parts.

One thing I did notice: a small warning triangle sits next to the account name in the top right with no label or tooltip explaining what it flags.

On my account, it appeared immediately after signup, which suggests it’s a billing or verification prompt, but it took a click into account settings to confirm that, rather than the dashboard surfacing it plainly.

3. Launching an Instance

Clicking ‘Add Server’ from the Flexible panel opens a single long form rather than a multi-step wizard. All choices are visible at once.

The decisions, in order:

• Select Application: WordPress, Optimized WordPress, Basic WordPress (No Cloudways Optimization), WooCommerce, WooCommerce with Optimized WordPress, or PHP
• Select Application Stack: Lightning Stack (Nginx, recommended) or Hybrid Stack (Apache/Nginx)
• Select Your Cloud Provider: DigitalOcean, Vultr, Linode, AWS, Google Cloud
• Select Your Database: MariaDB v10.11 (recommended) or MySQL v8.4
• Select Server Type: Basic, General Purpose, or CPU Optimized, each with a Premium/Standard toggle
• RAM slider: 1GB through 96GB depending on the type
• Server Location: dropdown, defaulted to New York

Two name fields follow, one for the server and one for the application, since Cloudways treats them as separate objects. A project selector lets you group servers from the start. Then ‘confirm’.

Confirming the launch does not go straight to provisioning. A modal appears offering five plugins:

A ‘Skip for Now’ link sits at the bottom left. I skipped it.

None of the four free plugins are relevant to a clean performance test, and two of them, the cookie consent tool and the accessibility checker, inject scripts on every page load. Installing either before running GTmetrix would mean measuring their overhead alongside Cloudways’ own stack, which tells the reader nothing useful about the hosting.

One click past that, and the server appeared in the list with a spinner and a quoted wait time of approximately 7 minutes.

It was ready in 5.

The single-page form approach is the right call here, and I appreciated it. Seeing all choices at once means I can make decisions in context rather than committing to one thing per screen and losing the view of what came before. The plugin upsell modal is the one interruption that broke the flow, and while none of the plugins would hurt a new user, the timing feels off.

Someone who just confirmed a server launch is not in the mood to evaluate marketing tools before the server has even started. The fact that the server came in under its own estimate, rather than over it, is worth calling out.

4. Server Management

Clicking the application from the dashboard opens directly to an Access Details screen.

Everything a developer needs on day one is visible on arrival, with no extra navigation required:

• Application URL: the live cloudways.cloud subdomain, with an external link icon
• WordPress Admin Panel URL: separate from the application URL
• WordPress credentials: username and password, both with copy buttons
• Database credentials: DB name, username, and password, plus a Launch Database Manager button
• Redis credentials: prefix, username, and password already populated, meaning object caching is already active

Below that: a Password Protection toggle for gating the site during development, a Public IP field, and an empty SSH/SFTP credentials table with an Add SFTP User button.

The left sidebar surfaces everything else under the application view:

• Staging Management
• Monitoring
• Application Security
• Domain Management
• Site Manager (labeled New)
• Cron Job Management
• SSL Certificate
• Backup and Restore
• Deployment via Git
• Application Settings
• Web Rules
• Cloudflare
• Migration Tools
• SafeUpdates

A server-level view is reachable via the breadcrumb at the top rather than a separate menu.

There is one friction point worth naming. A banner at the top of the Access Details page prompts me to migrate an existing website to Cloudways. I’d just finished building the server. The banner is aimed at the wrong user at the wrong time.

Everything I needed after launch was already on the screen without hunting. The database credentials being on the same page as the application credentials is convenient. Redis being pre-configured rather than requiring a separate setup step is a better default than most providers offer. The sidebar menu is long but logically ordered. The migration banner is a minor annoyance on an otherwise well-organized screen.

Overall Verdict on Ease of Use

Cloudways is one of the tidier managed hosting platforms I have put through this process. Sign-up is quick and genuinely requires no payment details upfront. The deploy form gives you full visibility of every configuration choice in one place rather than parceling them out across tabs.

The server was active in 5 minutes against a quoted 7, and the Access Details screen handed me database, Redis, and SSH starting points without making me dig.

The friction points I noted, an unlabeled account warning on the dashboard, an upsell modal between confirm and provision, and a migration banner on a brand new server, are real but none of them block anything.

A first-time user can go from signup to a running WordPress server in under 10 minutes without reading any documentation, and that is a meaningful benchmark that not every provider in this category clears.

Cloudways

Unparalleled speed, robust reliability, and tailored cloud hosting packages from Cloudwave are built with modern solutions in mind. Ideal for all web hosting needs, and extremely user-friendly.

Visit Cloudways

Level of Support

Cloudways offers three main support channels on all plans:

There is also a Community Center and a sales contact page offering a scheduled call option, but the latter is aimed at prospective customers rather than active accounts needing technical help. No direct phone support exists for standard users.

One thing worth flagging before anything else: support is tiered. My account ran on the Standard scope, and a banner inside the Support Center made clear that priority response times, application support, and server optimizations sit behind a paid upgrade. I tested what a standard paying customer gets, not a premium add-on.

Live Chat

The entry point is a tab on the right edge of every dashboard screen labeled “Need a Hand?”. Clicking it opens a widget with three things visible immediately: a message input, a live status notice (on the day I tested, it read “3 open incidents, updated Jul 18, 08:39 UTC”), and a list of suggested knowledge base articles below.

That status notice is a detail worth calling out as a positive. Most platforms bury incident reports on a separate status page. Cloudways surfaces it in the same widget you open to ask for help.

Clicking “Send us a message” starts the chat and routes you through the AI agent first.

The AI experience

The AI opened with two category buttons: Technical Help and Billing Help.

I chose Technical Help and got a second screen with 18 problem categories:

• DNS / Migration / Take website live
• SSL / Redirects / .htaccess
• SSH/SFTP
• Cloudflare
• SafeUpdates
• Backup & Restore
• Cloudways Copilot
• Cloning / Staging
• Slow website / Caching / Varnish / Breeze / CDN
• Website / Application Errors
• Server Down or High CPU Utilization
• Database / Redis / Memcached / Elasticsearch
• Monitoring / Bot alerts
• Disk Full / Disk Space
• Security / Abuse
• Malware / Site hacked
• NodeJS / NPM / Composer
• Cron jobs

None of those categories matched what I needed, so I selected “My problem is not listed above” and typed my own question. I asked whether a PHP extension unavailable through the dashboard’s PHP settings could be installed manually via SSH, or whether that had to go through the Cloudways team.

I also asked the SafeUpdates question directly: if the platform handles automated security patching, would a manually installed extension survive that process or get overwritten.

The AI replied within the same minute. Its answer on the first part was correct and confident: you cannot install PHP extensions via SSH yourself, and all extension requests go through the support team.

On the second part, the SafeUpdates question, it hedged: “the docs don’t mention it affecting PHP extensions, since extensions aren’t installed manually on the platform.” That is not a direct answer. It is the AI telling me it cannot find a contradiction in the documentation, which is a different thing entirely.

The AI’s category routing is well thought out, and the first part of the answer was accurate. The SafeUpdates response is where the limitation shows.

An AI that says “the docs don’t mention X” when asked whether X happens is telling you the limits of its search, not the limits of the platform. That distinction matters for a user making an infrastructure decision.

The human handoff

I asked to be connected to a human agent. The AI confirmed immediately and Sheeza joined the conversation one minute later.

Sheeza’s opening move was to ask which specific extension I needed, a reasonable question in a real support scenario, but slightly deflecting when the question is about process rather than a specific install request. I clarified I was testing for a review and rephrased around imagick and sodium as examples.

Her full response, timed at about two minutes after my clarification:

• The technical team handles the installation and checks for server conflicts
• The SLA for such requests is 12 hours, though senior engineers often respond faster
• Once installed, the extension is left untouched by automated platform processes

She also linked directly to the knowledge base article on base packages deployed with a Cloudways server, which showed she was cross-referencing rather than answering from memory alone.

Timeline summary:

Sheeza’s answer was the kind of response that actually moves things forward. The 12-hour SLA is honest, which I appreciated more than a vague “as soon as possible,” and the confirmation that manual installations survive automated maintenance was exactly what the question asked for.

The closing message (“thank you for being an amazing customer”) was a little over the top for a nine-minute technical chat, but that is a style choice, not a quality problem.

Knowledge Base

With the chat test done, I moved to the other side of the support picture: whether a user who prefers to find answers themselves, rather than waiting on an agent, has the materials to do that.

I went back to the Support Center and clicked “Go to Knowledge base” in the top right of the support articles section.

The top-level structure has nine collections:

Counting across all nine collections, there are over 330 articles, with a search bar at the top and a “New to Cloudways? Start here!” section below the grid pointing to six onboarding articles directly.

That is a solid library on paper, but article counts tell you nothing about whether the content inside is any good. I wanted to open one of the deeper categories, then an actual article, to see whether the writing holds up, whether the screenshots match the current interface, and whether the articles end somewhere useful or just stop.

I picked the Administering Server and Website collection, the largest in the library at 123 articles, on the assumption that this is where most active users will spend their time once they are past initial setup.

I opened “Find Your Application Folder Name on Cloudways with Step-by-Step Guide,” a routine task that a developer SSH-ing into their server for the first time would realistically need to look up.

Written by Syed Abuzar Mehdi and dated April 13, 2026, meaning it was updated after the platform’s recent interface changes, not carried over from a previous version of the dashboard.

The article runs four numbered steps, each paired with an annotated screenshot. The screenshots matched what I see in my own account, which is not always the case with hosting documentation that lags behind UI updates.

Each step is a single action, nothing is buried in a paragraph, and the article closes with a “Need Help?” block pointing back to chat and ticket support rather than just ending. A thumbs rating widget sits at the bottom.

I came away with a positive read on the knowledge base. The article I picked is a simple one by design, but it cleared the tests that matter: current screenshots, numbered steps instead of prose walls, and a useful exit rather than a dead end.

With 123 articles in that one collection alone, I couldn’t read them all, but the structure and maintenance standard on the one I did open gives me reasonable confidence the rest is kept to a similar standard.

Overall Verdict on Level of Support

The live chat experience held up well under a question that required a real answer rather than a documentation link. The AI handled the easy part correctly and showed its ceiling on the harder part.

The human handoff took one minute, and Sheeza’s response covered both parts of the question with a concrete SLA figure and a direct answer on automated maintenance.

The knowledge base is large, recently maintained, and organized in a way that makes it usable without a search query.

The one thing prospective customers should know going in: Standard scope is what you get by default, and priority response times cost extra.

Cloudways

Unparalleled speed, robust reliability, and tailored cloud hosting packages from Cloudwave are built with modern solutions in mind. Ideal for all web hosting needs, and extremely user-friendly.

Visit Cloudways

Cloudways Review: Bottom Line

Cloudways earns a clear recommendation for WordPress developers and agencies who want managed hosting without being locked into a single cloud provider. The 30-day test produced 100% uptime, a GTmetrix Performance average above 87%, and a CLS of zero every single day.

Redis and Nginx came pre-configured on a server that took five minutes to build. The platform earns its management fee by doing that work invisibly.

It is the right fit for developers managing client sites, agencies running multiple WordPress applications, and anyone who wants cloud infrastructure flexibility without AWS-level complexity. The five-provider choice is a genuine advantage for teams with specific geographic or compliance requirements.

It is not suited to users who need a refund safety net, the Terms of Service are explicit that paid fees are non-refundable. Windows hosting is unavailable, priority support costs extra, and PHP extension installs go through a support ticket with a 12-hour SLA rather than a terminal command. Know those limits going in and Cloudways rarely gives you a reason to look elsewhere.

Cloudways

4.6

537 user reviews

CA$15.49 /mo

Starting price

Visit Cloudways

Rating based on expert review

• User Friendly
  4.5
• Support
  4.6
• Features
  4.8
• Reliability
  4.8
• Pricing
  4.3