Every layer between your users and your service.
14 test types, run from agents you place anywhere, with automatic diagnosis on top. Here is what each one measures and why it matters.
Agent-to-server and agent-to-agent tests measure loss, latency and jitter, and trace the route hop by hop.
Multi-flow Paris traceroute
Several flows with their own source port or ICMP id reveal load-balanced (ECMP) branches; silent branches are not counted as loss.
Real vs. rate-limited loss
Loss counts only when it persists downstream, so routers that deprioritise traceroute replies don't trigger false alarms.
Hop networks and roles
Every hop mapped to its AS and role: your site, access ISP, transit, destination.
MPLS, tunnels, DSCP
MPLS labels (RFC 4950), routers hidden in pipe-mode tunnels, and where your DSCP marking gets re-written.
Path MTU
DF probes find the path MTU, fragmentation-needed messages and PMTU black holes.
Agent-to-agent
One-way latency (NTP-corrected), loss per direction and TCP throughput between your own sites.
Your prefixes watched live from public route collectors (RIPE RIS, RouteViews) and from your own routers over BMP.
Origin hijacks
Announcements of your prefixes by an AS you didn't list.
Sub-prefix hijacks
More-specifics of your space appearing anywhere.
Route leaks
Valley-free violations confirmed from at least two monitors, to keep noise out.
RPKI
Announcements that are RPKI-invalid for your prefixes.
Unexpected upstreams and reachability
New neighbours of your origin, and prefixes disappearing from the table.
Path-linked prefixes
The prefixes your tests' traffic crosses are watched too, so a BGP burst shows up next to the loss it caused.
Simple lookups against any resolver, and full traces that resolve from the root servers down.
Trace from the root
Every referral, CNAME and glue lookup with its timing: root, TLD, authoritative.
DNSSEC validation
The full chain of trust from the root anchors (KSK-2017 and KSK-2024) to the answer; bogus answers fail the run.
Broken delegations
Lame servers, glueless delegations and upward referrals called out.
From a single status check to a scripted journey in a real browser.
HTTP
Status, body checks and timing phases from every agent.
Multi-step API
Up to 10 steps sharing cookies, with variables, secrets and values extracted from earlier responses.
Page load
Headless Chromium with a cold cache: navigation timing, FCP, LCP, CLS, the waterfall, failed requests, JS errors.
Browser transactions
Log in, search, check out: scripted steps, each timed, with a screenshot when one fails.
FTP, FTPS, SFTP
Connect, handshake, login and transfer phases timed separately.
Know call quality before your users do.
Voice paths
RTP-sized packets at 50/s, marked EF, between two agents in both directions.
MOS and R-factor
ITU-T G.107 E-model per direction for G.711 and G.729, with jitter-buffer discards and burst loss.
SIP
OPTIONS and REGISTER over UDP, TCP or TLS with digest auth, retransmits counted.
Vantage points wherever your users and services are.
Install anywhere
One command on Linux (deb, rpm, binary), Docker, macOS and Windows; agents only make outbound HTTPS calls.
Launch in your AWS account
Pick regions in the dashboard; we start, enroll, update and clean up the instances.
Shared agents
Use our hosted agents in other regions, within fair-use limits.
Laptop agents
Endpoint agents also report Wi-Fi signal, gateway latency and loss, VPN and DNS server timing, and keep a day of data offline.
Auto-update
Agents update themselves (checksums verified); dashboard-launched ones are replaced without losing history.
Alerts that say what happened, sent where your team works.
Threshold and baseline rules
Fixed limits, or deviations from each agent's own 7-day normal (median and MAD).
Fewer false alarms
Sustained for N runs and seen by at least M agents before anything opens.
Channels
Slack, Microsoft Teams, PagerDuty, ServiceNow, email and webhooks, routed per rule or by severity.
Incidents
Acknowledge, notes, maintenance windows, and alerts to Slack, Teams, email, PagerDuty and ServiceNow.
Dashboards and snapshots
Build dashboards, and share a frozen page with a public link that expires.
See whose network it is, on your own traffic
Free during the beta. Install an agent in one command and your first test is live in minutes.