Lead service
Cloud infrastructure and platform operations
I design, build and run the infrastructure your business sits on, and then I keep running it.
Who this is for
- You have production systems on AWS, or somewhere you've outgrown, and nobody whose actual job is looking after them
- Your platform works but you don't know what happens if the person who built it stops answering
- Your cloud bill has grown faster than your business and nobody can tell you why
Who it isn't for
- Anyone who needs a named engineer on call around the clock for a bespoke platform
- Anyone looking to hand over a very large build in one piece
This is planned work, scoped and scheduled, rather than an on-call contract or a twelve month programme. If you need either, I'd rather say so now and point you at somebody who does it properly.
WordPress sites on a care plan are a different arrangement and are monitored around the clock. See hosting.
The problem
Infrastructure is easy to build and hard to run. Standing up a server is an afternoon. Keeping it patched, backed up, monitored, performing and secure for eight years is a job, and it's a job most companies your size can't justify hiring for.
So it doesn't get done. The instance keeps running because instances usually do. The backups are configured but nobody has restored one. The database gets slower by a percent a month and nobody notices until the year it doesn't fit in memory any more. The person who set it up has left and their documentation is a Slack thread.
That's the normal state of things. It isn't negligence, it's arithmetic. The work is real but it's nobody's full-time role.
What I do
- AWS architecture and management. Multi-server estates, sensible boundaries, things that can be rebuilt.
- Cloudflare. DNS, CDN, WAF, and the configuration that actually makes a difference rather than the tickbox version.
- Linux administration. The actual servers, hardened properly.
- Database performance. MySQL, MariaDB and Aurora, tuned by somebody who has been doing it since dBase III+.
- Backup and disaster recovery, including the part where we test the restore, which is the only part that counts.
- Security hardening and patching. Proactive, scheduled, documented.
- Migrations. Between providers, between architectures, out of things nobody wants to touch.
Cost reduction
Worth its own heading, because it's the easiest thing in infrastructure to justify: it pays for itself and you can measure it.
Most cloud estates carry real waste, and almost none of it is exotic. It's instances sized for a load test that ran in 2021, storage nobody deleted, data transfer routed the long way round, and three environments where two would do.
If your AWS bill is a number you flinch at and can't explain, that's a technical review, and it usually pays for itself in the first quarter.
How it works
Most engagements start with a technical review. I look at what you've got and give you written findings, prioritised, with costed options. Some people take that document and hand it to their own team, which is fine.
If you want me to do the work, it's quoted as a piece of work against an agreed scope rather than as time. Planned, by somebody who already knows your system.
Consultancy prices on this page exclude VAT.
What it runs on
- AWS
- Cloudflare
- Linux
- MySQL
- MariaDB
- Aurora
- GridPane
- Vultr
- DigitalOcean
- Patchstack
- Redis
- Postmark
Proof
Twenty-six years of keeping one platform up.
The infrastructure behind a multi-tenant learning platform that has served more than 150,000 students, through every provider migration, database growth curve and security change in that period.
That's the longest continuous operational record I can show you, and operations is what this page is selling.
Questions
The things people ask
How quickly do you respond?
Work here is scheduled in advance rather than reactive, so it doesn't carry a response window. The dates are agreed and that's the commitment. WordPress care plans are the exception and carry round-the-clock monitoring.
Do you take over infrastructure somebody else built?
Usually, yes. That's most of the work. The first step is a review so we both know what's actually there.
Who owns the accounts?
On this kind of work, you do. Your AWS or Cloudflare account, your servers, your DNS, all in your name and paid for by you. I have access rather than ownership and you can revoke it whenever you like.
That is not true of my WordPress hosting, where the server is mine and shared. Different arrangement, different page, and it says so plainly.
What if you're not available?
Everything documented, infrastructure as code where possible, and accounts in your name. The full answer is on the about page and I'd rather you read it before you hire me than after.
Do you work on site?
No. Everything I do is remote and always has been.
What size company do you work with?
Broadly ten to a hundred people. Big enough that the systems matter, small enough that hiring for this doesn't make sense.