Latency
Measure round-trip time from important user and office networks, not only from a cloud test node.
Network
Measure the route that matters to your users.
Network marketing numbers are not enough. Test latency and transfer performance from the actual regions and ISPs that matter to your application.
Quick answer
For the Georgia VPS line, the confirmed public network profile is 200 Mbps local Georgian connectivity and 10 Mbps international connectivity with unlimited monthly traffic. Test from the locations that matter to your users instead of treating one benchmark as universal truth.
Testing
What to measure
Packet loss
Persistent loss can harm interactive applications even when nominal bandwidth looks high.
Route
Traceroute or MTR can help identify where a path changes, but interpret results carefully because some hops deprioritise probes.
Real transfer
Use application-relevant transfers and several locations. One speed test should not be treated as universal performance.
Transparency
Looking Glass endpoints are published only when they are real
This package intentionally does not invent test IPs or benchmark URLs. Add production test endpoints in config.php once you have confirmed which SERVER1 network resources are intended for public testing.
Test IP
Publish an address designed for public diagnostics, not a customer or management interface.
Test file
Use a controlled public test object and monitor abuse or unexpected transfer cost.
Looking Glass
Expose only safe diagnostic functions and rate-limit them appropriately.
Practical interpretation
How to read the network profile correctly
Local and international connectivity values describe the public VPS service profile and should be interpreted in that context.
The published 200 Mbps local and 10 Mbps international profile helps set expectations for multi-tenant Georgia VPS service. It should not be compared directly with dedicated server port profiles or with generic advertising from unrelated providers.
For business-critical workloads, test from your actual audience regions, office networks or upstream systems. Latency, packet loss and route quality matter just as much as raw transfer speed.
Where public Looking Glass or test objects are not yet published, keep the copy factual and invite customers to ask for workload-based guidance rather than inventing diagnostics that do not exist.
FAQ
Questions people ask before ordering
A port speed is only one constraint. Internet performance also depends on congestion, peering/transit, remote networks, route choice, protocol behaviour and the application itself.
Publish or provide a test endpoint only when SERVER1 has designated one for public diagnostics. This site intentionally does not fabricate a placeholder address.
Choose by workload, not by guesswork.
If you know the resources you need, open the live catalogue. If not, send the workload and we will help narrow the choice.