Where your server actually runs

Today

VPSLatch runs on leased capacity in a single European facility, on AMD EPYC hosts with NVMe storage and a 1Gbps uplink per node. We are a small operation buying quality capacity rather than a network operator with our own racks, and pretending otherwise would be the sort of claim you could check.

That has a real consequence worth stating plainly: we do not control the datacentre. If the upstream facility has an outage, we are in the same position as you — waiting, chasing, and telling you what we learn as we learn it.

What we will commit to

• Nodes are not oversold on memory. Allocated RAM is allocated.
• Storage is local NVMe. There is no network storage layer to fall over.
• You get told about planned maintenance before it happens, not after.
• Incidents get an honest post-mortem, including the ones that were our fault.

What we will not claim

• We do not claim a tier rating for a building we do not run.
• We do not claim DDoS protection beyond what the upstream provides, which is basic volumetric filtering.
• We do not claim 99.99%. Our SLA says 99.9% and explains what happens when we miss it.
• We do not publish a network map with lines to cities we have no presence in.

Backups

We do not back up your server. Node-level redundancy protects against a disk failing; it does not protect against you running the wrong rm, and it is not a backup. Take your own, off this machine. If you want us to take a manual snapshot before something risky, open a ticket and we will.

Growing from here

More locations, self-service reinstall and automated snapshots are the obvious next steps. We will list them on this page when they exist and not before. If a second region would decide your order, say so — that is genuinely how we will prioritise.