Immediate
Infrastructure is down
A machine is unreachable, a switch port is dark, a facility has an open incident. These are escalated the moment they are seen, at any hour, because they are almost always ours and almost never yours.
- Host unreachable from every direction
- Hardware fault or failed disk
- Routing withdrawn for your address
Same day
Something is degraded
The service works but not properly. These need a person to look at both sides, and that person is the one who administers the hardware.
- Packet loss or asymmetric routing
- A disk reporting errors but still serving
- Reverse DNS, addressing and allocation
Working day
Everything else
Questions, changes and billing. Nothing here is less important — it is just work that is not on fire, and treating it as if it were is how the first two bands stop being fast.
- Reinstalls and image requests
- Invoices, cancellations and upgrades
- Pre-sales and configuration questions
These describe how the queue is worked, not a contract. We do not publish a response-time guarantee we have not signed. If you need one in writing for a specific service, ask before ordering and we will tell you plainly whether we can offer it — some accounts have exactly that, agreed in writing as extended round-the-clock cover, and the case studies say so on the page where the account is named. It is still this queue and still these people; what it changes is when we are reachable and how fast a fault is escalated.
