Service
AWS and cloud infrastructure
AWS consulting and cloud infrastructure: setting it up, migrating onto it, and working out why the bill went up.
AWS has a service for everything and no opinion about which ones you need. Accounts drift. Instances stay sized for a launch that never came, volumes sit attached to nothing, and the bill grows faster than the business does.
We set it up around what your application actually does. Compute, database, storage, CDN, DNS and certificates, with separate access for each person and each service instead of one key passed around the office. Deploys get automated so releases stop being an event. Monitoring goes in so problems reach a person.
We also read bills. Line by line, tagged, with a note on what each thing is for and what happens if it goes. The biggest saving is usually something dull.
What you get
An architecture that fits the workload
Sized for what the application does today, with a note on what to change when traffic doubles, rather than paying for that capacity now.
Migration without a shutdown
Onto AWS, between regions, or back off it. The old environment stays live until the new one has handled real traffic.
Access under control
A role per person and per service, keys rotated, root locked away. Most cloud incidents start with credentials rather than code.
A bill you can explain
Every recurring line identified and tagged, with the ones that can go safely marked as such.
Questions about aws and cloud infrastructure
Whose AWS account is it?
Yours. You pay Amazon directly. We're not a reseller and we don't add anything to your bill. We work inside your account with the access you give us, and you can take it away whenever you want.
Is AWS right for us?
Often it isn't. A single application with steady traffic usually runs cheaper and simpler on a dedicated server. AWS earns its money when you need managed databases, traffic that moves around a lot, or more than one region.
Can you reduce what we're already paying?
Usually. Idle load balancers, oversized instances, old snapshots and volumes nobody deleted are the usual suspects. You get the findings before we change anything.