Cloud & infrastructure
Infrastructure run by people who've been paged at 3am
Migration, DevOps and managed cloud from a team that operates its own production hosting platform — so the advice comes from running things, not from a certification.
Cloud runs in your accounts, billed at your vendor's price. No markup.
What this practice covers
Cloud and infrastructure work is everything between your code and your customers: where it runs, how it gets there, how you know it's healthy, and what happens at 3am when it isn't. Migration, automation, monitoring, backups, security patching and cost control.
We came to this from an unusual direction. HostMentor started as a hosting and domain business, which means we were running production infrastructure, answering support tickets and handling incidents long before we were advising anyone else about it. That's a different foundation from a consultancy that learned cloud from a certification path.
The practical consequence is that we optimise for operability rather than for architecture-diagram elegance. A simpler system your team can actually run beats a sophisticated one that needs us forever.
Managed hosting, cloud, or your own servers?
A genuine decision, and one where most companies overspend by picking the most impressive option rather than the right one.
| Managed hosting | Public cloud | Dedicated / on-prem | |
|---|---|---|---|
| Best for | Websites and standard applications with predictable load | Variable load, many services, rapid scaling | Fixed heavy workloads, or strict residency rules |
| You manage | Your application only | Everything above the hypervisor | Everything, including hardware lifecycle |
| Cost shape | Flat monthly, predictable | Usage-based — cheap small, expensive careless | High upfront, low marginal |
| Scaling | Plan upgrades | Elastic, near-instant | Procurement cycle |
| Ops burden | Lowest — it's our job | Highest without automation | High, plus hardware |
| Typical mistake | Outgrowing it and not noticing | Lift-and-shift, then a shocking bill | Buying for peak load you hit twice a year |
Plenty of workloads that get moved to public cloud would run better and cheaper on managed hosting. We sell both, so we have no reason to push you either way.
What we do
Six services across moving to the cloud, running it, and keeping the bill sane.
Cloud migration
Moving applications from on-premise or another provider with a tested rollback at every stage. We migrate in slices rather than one weekend cutover with everything riding on it.
Infrastructure as code
Your environment defined in Terraform and reproducible from scratch. Ends the 'nobody knows how staging was built' problem permanently.
CI/CD pipelines
Automated build, test and deploy with preview environments per change. Deployment becomes routine rather than an event people schedule around.
Monitoring & incident response
Metrics, logs, tracing and alerts that fire on symptoms customers feel rather than on every CPU spike. Plus a written runbook per alert.
Backup & disaster recovery
Automated backups with restores actually tested on a schedule. An untested backup is a hypothesis, and we've seen too many fail at the worst moment.
Cost optimisation
Right-sizing, reserved capacity, storage lifecycle and finding the things nobody turned off. Frequently pays for the engagement in the first quarter.
How we migrate without a bad weekend
Five stages, each reversible. Nothing moves until the thing before it is proven.
Inventory
What actually runs, what depends on what, and which of it nobody has touched in three years. This stage routinely finds servers no one can account for.
Target design
The destination architecture, sized against real measured usage rather than the current over-provisioned shape, with a costed estimate.
Pilot
One non-critical workload moved end to end. Proves the approach, the automation and the runbook while the stakes are low.
Migrate in waves
Workloads move in dependency order, each with a rehearsed rollback. Data syncs continuously so cutover is a DNS change, not a restore.
Optimise & hand over
Right-size against real post-migration usage, tune alerts, and hand over runbooks and access — plus a support window while your team settles.
We keep the old environment running until you've had a full billing cycle on the new one. Rushing decommission is how migrations become incidents.
Why cloud bills surprise people
Four causes we look for first. Between them they account for most of the waste we find.
Lift-and-shift sizing
On-premise servers were sized for peak load plus headroom, and that shape gets copied into instances billed by the hour. Right-sizing against measured usage is usually the single biggest saving.
Data transfer
Egress and cross-zone traffic are the charges nobody models in advance. A chatty architecture across availability zones can cost more than the compute it connects.
Nothing ever gets deleted
Orphaned volumes, old snapshots, idle load balancers and environments from a project that ended last year. Storage lifecycle rules fix this permanently.
No cost ownership
When the bill isn't attributed to teams, nobody optimises. Tagging and per-service cost reporting turns an abstract number into someone's responsibility.
Or just buy hosting
Not every workload needs a cloud engagement. Our managed hosting products are priced, self-serve and run on the same infrastructure we operate for enterprise clients.
Web Hosting
Fast NVMe shared hosting
Plans & pricingWordPress Hosting
Managed, auto-updated WordPress
Plans & pricingCloud & VPS
Scalable servers with root access
Plans & pricingBusiness Email
Email on your own domain
Plans & pricingDomains
Search, register & transfer domains
Plans & pricingFree migration, daily backups and free SSL are included on every plan — and the same engineers support both.
FAQ
Cloud & infrastructure questions
Which cloud should we use?
Usually whichever your team already knows, unless a specific requirement points elsewhere — data residency, a managed service only one provider offers, or existing credits. The differences matter far less than most comparisons suggest, and we don't resell any of them, so the recommendation is unbiased.
Do you resell cloud with a markup?
No. Cloud accounts are opened in your name and billed directly to you by the vendor, so you see the real invoice and keep any credits or discounts. We charge for our work, not for passing through infrastructure. Our own hosting products are the exception — those are our service, priced as such.
How long does a migration take?
A single application is typically two to six weeks including the pilot. A full estate of dozens of workloads runs three to nine months, moved in waves. We deliberately don't compress this — the schedule pressure is where migrations go wrong.
What if we want to leave you afterwards?
Everything is in your accounts and defined in Terraform in your repository, so there's nothing to hand back — you already have it. Handover is runbooks and a knowledge-transfer window rather than a data extraction exercise.
Can you take over infrastructure someone else built?
Yes, and it's a common engagement. We start with an audit: what exists, what's exposed, what's not backed up, and what it would cost to bring to a supportable standard. You get that assessment as a document whether or not you continue with us.
Can you meet our data residency requirements?
Yes. We design to your residency requirements from the start and document processing locations and data flows as part of delivery, since you'll be asked for that evidence at audit rather than at design time.
Related services
What's running your business today?
Tell us where it runs now and what's painful about it. We'll come back with an honest read — including when the answer is to leave it where it is.