Talk to us

Case studies

What was wrong,
and what we did about it.

Three real jobs, told plainly. Sector and scale are real; names are not, because our clients' infrastructure is their business rather than our marketing material.

Three case studies

Cinema Microsoft Azure Subscriptions, payments and configuration

The cinemas kept going offline, and the cause was not the infrastructure. It was a payment that had not gone through.

A cinema chain ran its systems on Microsoft Azure and was losing service repeatedly. Some of it came from configuration that had drifted over time. Most of it came from something more mundane: subscription payments were not reaching Microsoft when they needed to, and services were being suspended as a result.

Nobody inside the organisation owned the Azure account end to end. Payment sat with one person, configuration with another, and the outages lived in the gap between them, which is why they kept recurring even after each one was individually fixed.

We took the whole thing over: the subscriptions, the payment schedule, and the configuration underneath. Renewals are paid on time because paying them is now our job rather than an item on somebody's list, and the environment has been tidied so that when something does go wrong it degrades quietly instead of taking screens down.

Outages caused by payment delays
None since
Their time spent on renewals
None
Azure environment
Managed end to end
Media VPS to dedicated server Migration, then ongoing management

They had outgrown the VPS. What they could not afford was to be dark while moving off it.

A media organisation had grown past what its virtual server could carry. The answer was obvious: move to a dedicated machine. But the risk was not acceptable to them. An audience that finds a publisher offline reads whoever is still up, and some of them do not come back.

So we did not move them onto anything. We built the dedicated server alongside the VPS while the VPS carried on serving, verified it under real conditions, and only then pointed traffic at it. If the new server had not behaved, nothing would have changed for their readers.

From the decision to being live on dedicated hardware took under 24 hours. We have run the server since.

Downtime during the move
None
Decision to live
Under 24 hours
Since then
Managed by us
Domain portfolio Multiple registrars Consolidation, then ongoing management

Nobody could say for certain how many domains the organisation owned, or who was supposed to be renewing them.

The portfolio had accumulated the way most portfolios do: a domain registered here for a campaign, another there when someone needed it quickly, spread across several registrars over a number of years. Some were country extensions, which carry their own renewal windows and their own awkward administrative processes, and which very few people enjoy dealing with.

Renewals had already started slipping. That is the part that matters: an expired domain does not degrade gracefully. The website goes blank and the mail stops, and recovering a name after it has lapsed is considerably harder and more expensive than renewing it would have been.

We inventoried the whole estate, brought every domain under our management, and took over the renewals. The client now receives a single invoice covering all of them, on one date, and has not had to think about a renewal since.

Renewals missed since
None
Registrars to deal with
One
Invoices to approve
One

Worth explaining

Why there are no client names

Publishing that a particular organisation runs on infrastructure we manage tells anyone interested where to start looking. For most of our clients that is a small risk, and for a few it is not a small one. So we describe the work and leave the names out.

If you are weighing us up and want to speak to somebody we already work with, ask. We will approach a client whose situation resembles yours and see whether they are willing. A reference given voluntarily is worth more than a logo we put on a page without asking.

Any of that sound familiar?

Most of this work started with someone describing a mess they had inherited. If that is roughly where you are, it is a normal place to begin.