Hosting & Cloud Instance Management

Hosting that quietly never breaks.

The server that holds your website, your forms, and your customers’ first impression. We provision it, tune it, harden it, and watch it — so the only people who notice the hosting are the ones who never have to think about it.

Sub-techniques covered · VPS · Cloud Instances · Managed Hosting · Server Configuration · Capacity Planning · Hosting Migration · Provider Selection · LNMP/LAMP
01 — What’s Included

Six hosting
responsibilities. One owner.

Hosting is rarely a single decision. It is a sequence — choose the provider, size the instance, configure the stack, harden the surface, plan for growth, and migrate cleanly when the original choice no longer fits. Every line below is a separate piece of work that we own end to end.

This is startup-grade infrastructure for businesses with one to fifty employees. Not an enterprise NOC. Not a cookie-cutter shared host. Honest, sized-to-fit hosting on the providers that actually serve growing companies.

N° 01

Provider Selection

First decision

The right host for a Sydney boutique law firm is not the right host for a cross-border e-commerce store with customers in Vancouver and Shanghai. We help you pick the provider that fits — Hetzner for value-driven European workloads, DigitalOcean and Linode for predictable VPS pricing, AWS Lightsail for teams already inside the AWS ecosystem, Alibaba Cloud Hong Kong for traffic that needs to reach mainland China without the Great Firewall tax, or a managed WordPress host like Cloudways or Kinsta when you want the convenience and the budget allows it. We do not take referral kickbacks, and we will tell you when the cheapest option is the right one.

N° 02

VPS & Cloud Instance Setup

Foundation

Provisioning a virtual private server or cloud instance from a fresh image and turning it into something production-ready. Operating-system selection, kernel tuning where it matters, swap configuration, time-zone and locale, automatic security updates, firewall rules, SSH-key-only access, fail2ban, and a non-root deploy user. We use LNMP (Linux, Nginx, MySQL/MariaDB, PHP) by default for WordPress and PHP workloads, with LAMP available where Apache is genuinely needed. Every server we hand over has a one-page reference document that explains what is installed, why, and how to log in safely.

N° 03

Server Configuration & Tuning

Performance

A fresh server is the starting line, not the finish. We tune the web server (Nginx worker counts, FastCGI cache, gzip and brotli compression, HTTP/2), the application runtime (PHP-FPM pool sizing, OPcache configuration, realpath cache, memory limits matched to your actual workload), and the database (MariaDB or MySQL buffer pool, query cache where appropriate, slow-query logging). The same droplet that ships at 200 ms under default settings will serve under 80 ms once the configuration is honest about your traffic. None of this is exotic; it is a checklist applied by someone who has done it many times.

N° 04

Managed Hosting Oversight

Vendor liaison

If you are happier on a managed platform — Cloudways, WP Engine, Kinsta, SiteGround, Pantheon — we manage the platform on your behalf rather than rebuilding from scratch. That means setting environment variables correctly, configuring caching layers and CDN settings, handling staging-to-production deployments, raising and chasing support tickets in plain English, and reviewing the monthly bill against actual usage. Managed hosting is sold as “set and forget” — in practice, every site benefits from someone reading the dashboard properly once a month.

N° 05

Capacity Planning

Sizing

The two most common hosting mistakes are paying for a server twice as large as the workload requires, and running on a server half the size the workload actually needs. We monitor CPU, RAM, disk I/O, and bandwidth against your real traffic and recommend a right-sized configuration — including the moment when a single VPS should grow into a small fleet, when MySQL should move to its own instance, or when a managed object storage bucket should take static assets off the application server. Capacity reviews happen quarterly on retainer engagements and on demand for project work.

N° 06

Hosting Migration

Without downtime

When the current host is wrong — slow, oversold, unresponsive, or simply too expensive for what you get — we migrate without losing visitors, email, or rankings. The standard migration is a parallel build: spin up the new environment, replicate the application and database, point a low-TTL DNS record at the new server during a quiet window, verify, then decommission the old. For larger sites we use rsync deltas, database replication, and read-only freeze windows to keep the transition under a few minutes of perceived disruption. We have done this for WordPress, WooCommerce, and custom PHP applications across at least a dozen provider pairings.

02 — Our Approach

Right-sized.
Hardened. Watched.

Reliability is mostly the product of doing four boring things consistently. We do not promise an enterprise data centre — we promise the small set of disciplines that, applied honestly to a startup-grade host, are the actual difference between “just works” and “why is the site down again”.

i

Sized to your real load

We measure before we recommend. A week of CPU, RAM, and disk-I/O graphs from the existing server (or a controlled staging build) tells us more than any sales pitch from a hosting provider. The goal is a configuration that runs comfortably at typical load, has headroom for the next twelve months of growth, and is not paying for capacity nobody is using. Right-sizing is an honest review, not an upsell pipeline.

ii

Hardened from day one

Every server we provision ships with key-only SSH, firewall rules locked to required ports, fail2ban tuned for real attack patterns, automatic security updates for the operating system, and a documented non-root deploy user. We are not running a SOC-2 audit; we are closing the doors that get rattled within an hour of any new IPv4 address coming online. The baseline is the same on a $5 droplet as on a $500 instance, because attackers do not differentiate.

iii

Documented for handover

The server reference for every hosting engagement lives in a shared document — provider, region, instance size, OS, stack versions, deploy paths, log locations, cron schedule, backup destination, and the monitoring endpoints that watch it. If we disappeared tomorrow, any competent technical person could log in, read the document, and be productive within an hour. This is the opposite of the “trust me, it just works” black box that growing businesses get stuck inside.

iv

Watched continuously

Provisioning is one event. Reliability is the months that follow. Every server we hand over is connected to uptime monitoring, resource-usage alerts, certificate-expiry watches, and disk-fill warnings — so the noisy first sign of trouble is an email to us, not a frustrated message from a customer. On retainer engagements we review the monitoring dashboard weekly. On project engagements we hand it over with a one-page guide on what each alert means.

03 — Who It’s For

Growing businesses,
without a sysadmin.

Hosting work fits best for small and mid-sized teams who have already outgrown the cheapest shared plan but are nowhere near the size that justifies a full-time infrastructure hire. The right help, sized correctly, costs a fraction of either extreme.

A handful of familiar profiles where startup-grade hosting is the practical fix.

  • i WordPress sites running on a VPS that nobody set up properlyThe droplet is fine; the configuration is generic. We harden it, tune the stack, document the surface, and the same hardware suddenly serves twice the traffic at half the response time.
  • ii Businesses paying enterprise hosting prices for startup workloadsThe site receives a few thousand visits a day, but the bill suggests an Amazon-scale operation. We migrate to right-sized infrastructure and the savings often pay for the engagement in the first quarter.
  • iii Cross-border operations needing reliable mainland-China accessAustralian or Canadian businesses serving Chinese-speaking customers run into the Great Firewall on every page load. Alibaba Cloud Hong Kong with a properly configured CDN typically resolves it.
  • iv E-commerce stores where the checkout slows down at peak hoursThe pattern is almost always under-sized PHP-FPM workers, an exhausted MySQL connection pool, or an overloaded shared CPU. The fix is rarely a bigger server — it is a better-tuned one, sometimes paired with object caching.
  • v Teams inheriting infrastructure from a previous agencyNobody on staff knows how it was configured, the access credentials are scattered across three former employees’ inboxes, and the documentation is a single password-protected zip file. We audit, document, and take ownership without rebuilding from scratch.

We do not host sites on infrastructure we own and rent back to you, and we do not lock anything we configure to our presence. The hosting account belongs to your business, the credentials are yours, and the server reference document we maintain is designed so any competent technical person could take over without a month-long handover. That is the standard — anything less would be vendor lock-in dressed up as a service.

04 — A complimentary report

Curious how Google sees your site?

Send us your URL. We’ll send back a Premium SEO Report, prepared by hand, within 48 hours — domain authority, keyword rankings, backlinks, competitor gap, and the quick wins worth chasing first. Hosting often sits behind the technical findings.

No sales call required.

Reliability is the work nobody notices. The best server is the one your customers never have a reason to think about.
— The Aureole Practice —
05 — Frequently Asked

Questions we get
asked about hosting.

If your situation is not covered here, the contact link at the foot of the page goes straight to the person who would answer it. No ticket queues, no funnels.

i Which hosting providers do you actually use?
For VPS and cloud-instance work we most often use Hetzner, DigitalOcean, Linode (now Akamai), and AWS Lightsail. For sites that need reliable mainland-China access we use Alibaba Cloud Hong Kong. For managed WordPress hosting we work most often with Cloudways, Kinsta, and WP Engine. We are not tied to any of them — the choice is driven by your traffic, your geography, your team’s familiarity, and your budget. We do not collect referral fees, so the recommendation is honest.
ii Do you offer enterprise-grade managed hosting with a 24×7 NOC?
No, and we will be straightforward about it. Our hosting work is scoped for businesses with one to fifty employees — startup-grade infrastructure with proactive monitoring and best-effort response during business hours. If you genuinely need a 24×7 network operations centre with a guaranteed sub-fifteen-minute response time, the right provider is a managed-services firm with that operational footprint. For most growing businesses, the practical alternative is a well-configured VPS with monitoring that prevents most outages from happening in the first place — and we are happy to set that up.
iii Can you migrate our site without downtime?
In nearly every case, yes — though “zero downtime” usually means a few minutes of read-only state during DNS cutover. The standard playbook is a parallel build: provision the new environment, replicate the database and files, lower DNS TTL forty-eight hours in advance, perform a final delta sync during a quiet window, switch the DNS record, and decommission the old environment a week later. For sites with large databases we use replication; for static and low-traffic sites we can usually keep the perceived disruption under sixty seconds.
iv We need our site to load quickly in mainland China — can you help?
Yes. The reliable approach for businesses without an ICP filing is to host on Alibaba Cloud Hong Kong (or a similar Hong Kong–based instance) and pair it with a CDN that has acceptable mainland reach. This avoids the cross-Pacific latency that makes Australian and Canadian hosts painful for Chinese visitors, and it does so without the complexity of a mainland ICP filing. Cloudflare’s mainland coverage has improved but is still inconsistent without a paid Enterprise plan, so we generally recommend a Chinese-friendly CDN or origin in Hong Kong as the practical compromise.
v How do we know we are paying for the right size of server?
A capacity review answers it directly. We pull a week or two of CPU, RAM, disk-I/O, and bandwidth graphs from your provider’s dashboard and your application logs, compare them against the instance specifications, and report back with one of three findings — under-sized, right-sized, or over-sized. About a third of the audits we run end with a recommendation to downsize. We are looking for the configuration that runs comfortably at typical load with headroom for growth, not the one that maximises the monthly invoice.
vi What happens if our hosting account is in someone else’s name?
First step is to take ownership. We help you create a hosting account in your business’s name, transfer the existing instance or migrate to fresh infrastructure, and lock down billing and access so that no future agency, contractor, or former employee holds the keys to your server. This is one of the most common starting points for new hosting engagements, and it is well worth doing properly the first time — the cost of recovering an account that is genuinely lost is significant.
The Invitation

Ready for hosting
that just works?

Tell us where the site lives today — provider, plan, the symptoms you are seeing — and we’ll respond within one business day with a clear assessment, an honest recommendation, and a plan that fits your scale.

Mon–Fri · 9–6 PT support@aureoleintelligence.com Reply within 1 business day