Multi-location probe (Looking Glass)
Run a check from several vantage points at once — our backend plus every remote node — and compare the results side by side. Ping shows latency, packet loss and jitter per region; traceroute and MTR show how the route differs; reachability, HTTP and DNS reveal regional outages, routing problems and geo-blocking. The target's network (ASN, prefix, country) is identified up front. More locations are added over time, and you can enroll your own servers as nodes.
Probe network
Vantage points are always-on nodes in different cities and networks. Each one runs the same checks from where it sits, so you can see how a target looks from around the world — not just from one place. 1 location · 1 online.
Sponsor a vantage point
We're growing the network and would love hosting providers to help. Donate a small, always-on VPS in a city we don't cover yet and it becomes a public probe node — credited to you right here, on the map and in this list, so everyone who runs a check from that location sees who made it possible. A modest box is plenty; we handle setup and keep it lightweight.
Offer a location → info@sysglass.comTCP-connect latency, loss and jitter. 0 remote nodes online; backend in Vilnius. More locations are added over time.
What you can run from each location
- Ping — opens a real TCP connection to the target (port 443 by default, or one you set) and measures round-trip latency, loss and jitter for every vantage.
- MTR — probes every hop on the path repeatedly and reports per-hop packet loss and latency, so you can see exactly where a path degrades.
- Traceroute — traces the routing path from each node to the target, hop by hop, to compare how regions reach it.
- HTTP — fetches status and headers from each location to surface geo-blocking, region-specific errors or CDN edge differences.
- DNS — resolves the name from each vantage, exposing geo-DNS, split-horizon and anycast answers that differ by region.
- Target network identified up front — the ASN, prefix and country/operator of the target are shown before the comparison, via ASN and IP geolocation lookups.
How to read the comparison
- Each row is one vantage point — the SysGlass backend in Vilnius plus every online remote node. The locations currently online are listed under the input.
- In the ping table, Loss is the share of connection attempts that failed (0% is healthy) and Sent is how many were tried.
- Last / Avg / Best / Worst are round-trip times in milliseconds; a node closer to the target should report a lower average.
- StDev is jitter — a large Best-to-Worst spread or a high StDev means an unstable path even when the average looks fine.
- A vantage that could not connect is flagged in red; if only some regions fail, the problem is regional rather than with the host itself.
- Different routes, HTTP statuses or DNS answers between regions point to asymmetric routing, geo-blocking or geo-DNS/anycast — not necessarily an outage.
Frequently asked questions
What is a multi-location Looking Glass?
A Looking Glass runs the same network check from several independent vantage points — the SysGlass backend in Vilnius plus every online remote node — and shows the results side by side. Because each location sits on a different network and in a different region, comparing them reveals problems a single test would miss: regional outages, asymmetric routing, geo-DNS differences and geo-blocking.
Which checks can I run from every location?
Five vantage-dependent checks: Ping (TCP-connect latency, loss and jitter), MTR (per-hop loss and latency along the path), Traceroute (the full routing path from each node), HTTP (status and headers, to spot geo-blocking) and DNS (resolution, where geo-DNS and anycast can differ). Checks whose answer is the same everywhere — WHOIS, SSL, mail authentication — are intentionally left out because one location is enough.
Is the Ping here ICMP or TCP?
It is a TCP-connect ping: each vantage opens a real TCP connection to the target (port 443 by default, or the port you specify) and times the handshake. Many hosts and firewalls drop ICMP echo but still accept TCP, so a TCP ping is a more reliable signal of whether a service is actually reachable. MTR and Traceroute use the path-tracing probes appropriate to each node.
Why do results differ between locations?
Latency grows with distance, so a node closer to the target reports a lower round-trip time. Differences in loss, route, HTTP status or DNS answer point to something more specific — a regional network problem, an anycast or geo-DNS setup that serves different IPs per region, or a firewall/CDN that blocks some countries. Side-by-side rows make those patterns obvious.
How do I read the per-region ping table?
Each row is one vantage. Loss is the share of connection attempts that failed (0% is healthy). Sent is the number of attempts; Last, Avg, Best and Worst are the round-trip times in milliseconds; and StDev is the jitter. A large spread between Best and Worst, or a high StDev, means an unstable path even when the average looks fine, and a row that could not connect is flagged in red.
Where are the vantage points, and can I add my own?
The home vantage is the SysGlass backend in Vilnius (Lithuania); every online remote node adds another location, and the list of online nodes is shown under the input. More locations are added over time, and you can enroll your own servers as nodes so checks also run from your own infrastructure. The target's network — ASN, prefix and country — is identified at the top before the comparison.