Global Watch Network

One location sees green.
Three continents see nothing.

Your monitoring says “all clear” because it only checks from one place. Pulse Stack checks from 10+ locations simultaneously so regional outages, CDN failures, and routing problems never go unnoticed.

LIVE: 10/10 locations operational
10 locations

NYC

32ms

SFO

45ms

YYZ

38ms

LHR

28ms

FRA

31ms

SIN

89ms

TYO

78ms

SYD

112ms

BOM

105ms

GRU

95ms

Multi-location checkexample.com
NYC
200 OK32ms
SFO
200 OK45ms
YYZ
200 OK38ms
LHR
200 OK28ms
FRA
200 OK31ms
SIN
200 OK89ms
TYO
200 OK78ms
SYD
200 OK112ms
BOM
200 OK105ms
GRU
200 OK95ms
10/10 LOCATIONS HEALTHY

How multi-location checks work

10 probes fire at once. The truth is in the consensus.

Every time a check runs, Pulse Stack sends requests from all enabled locations simultaneously. Each probe connects independently through its own network path, DNS resolver, and ISP routing. The responses come back at different speeds based on geography and network conditions.

If 9 locations get a 200 OK and one gets a timeout, that single failure is noise. If 4 locations in Asia all fail while the rest succeed, you have a genuine regional problem. The consensus algorithm separates signal from noise automatically, which means your alert channels only fire when there is a real issue.

This works across every monitor type: HTTP uptime checks, ICMP ping, TCP port scans, and API health checks all run from multiple locations by default.

False alarm elimination

Drag the slider. Watch the verdict change.

The consensus algorithm adapts its verdict based on how many locations report failure and which ones. A single-location failure is almost certainly a false positive caused by a local network issue at the monitoring node. Three failures in the same region point to a CDN or routing problem.

Teams running single-location monitoring get woken up at 3am for false alarms. They start ignoring alerts. When a real outage hits, the notification sits unread in a channel full of noise. Consensus-based alerting keeps your response sharp because every alert means something.

Set your own thresholds for how many locations must fail before an alert fires. Combine with incident management to automatically create incidents when the consensus verdict changes, and push updates to your status page so customers know what is happening.

Consensus simulator

0/10 failing
NYC
SFO
YYZ
LHR
FRA
SIN
TYO
SYD
BOM
GRU
0 failures10 failures

ALL CLEAR

Every location confirms your service is reachable. No action needed.

Response time by location

NYCNew York
32ms
SFOSan Francisco
45ms
YYZToronto
38ms
LHRLondon
28ms
FRAFrankfurt
31ms
SINSingapore
89ms
TYOTokyo
78ms
SYDSydney
112ms
BOMMumbai
105ms
GRUSão Paulo
95ms
<50ms
50-100ms
>100ms

Geographic performance

Know exactly where your service is slow

A server in London naturally responds faster to Frankfurt (31ms) than to Sydney (112ms). That is physics. But if London-to-Sydney suddenly jumps to 400ms, something changed: a routing issue, a CDN edge failure, or a provider problem that only affects traffic crossing certain network paths.

Sort by fastest or slowest to identify which regions experience the best and worst performance. Track these numbers over time through the response time dashboard to spot trends before they become outages.

If your CDN contract promises sub-100ms delivery worldwide, this data gives you the proof. Combine with DNS monitoring to separate DNS resolution delays from actual server response time. Use the SLA calculator to convert uptime percentages into real downtime minutes per region.

Start monitoring from 10+ locations today

Every monitor gets multi-location checks by default. Free for 50 monitors.

Start free

What multi-location monitoring catches that single-location misses

Three real-world scenarios where checking from one place gives you a false sense of security.

CDN edge failures are invisible from one location

Your CDN caches content at edge locations worldwide. When a CDN edge in Singapore stops serving cached content, response times spike for users in that region. But your monitoring server in Virginia sees perfectly normal performance because it hits a different edge.

Multi-location monitoring detects CDN edge failures by comparing response times across regions. A sudden 3x latency increase from one location while others stay flat points directly to a cache miss or edge outage. Pair with content monitoring to verify the cached content is actually correct, not just fast.

CDN edge status

LHR
HIT28ms
FRA
HIT31ms
NYC
HIT32ms
SIN
MISS340ms
TYO
MISS290ms

Geo-DNS misrouting goes undetected for weeks

Geo-DNS routes users to the nearest server based on their location. When it misconfigures, users in Europe might hit your US servers, adding 100ms of latency. Or worse, they get routed to a server that has been decommissioned.

Monitoring from each region verifies that geo-DNS resolves correctly everywhere. If Frankfurt checks start routing to a US IP instead of an EU IP, the latency jump shows up immediately. Add DNS record monitoring to track the actual resolution results per region.

DNS resolution by location

London185.23.108.2EU-WEST
Frankfurt185.23.108.2EU-WEST
New York45.77.64.12US-EAST
Tokyo45.77.64.12US-EAST
Singapore45.77.64.12US-EAST

Asia-Pacific resolving to US servers instead of APAC edge

ISP routing failures affect some users, not all

Internet routing is not a single path. Traffic from Tokyo to your London server passes through multiple autonomous systems, peering points, and undersea cables. When one link in that chain fails, only traffic from specific origins is affected. Your server looks healthy from its own network.

Multi-location monitoring reveals routing-specific failures that are invisible from your data centre. If your server responds to all locations except those routing through a specific transit provider, the pattern tells you exactly where to look. Combine with SSL certificate monitoring and TCP connection checks for full connection verification across every network path.

Connection status

NYCConnected
LHRConnected
FRAConnected
BOMTimeout at hop 9
SYDTimeout at hop 11

Simple pricing. Start free.

50 monitors free forever. Upgrade when you need faster checks or more capacity.

Free

£0forever

  • 50 monitors
  • 3-min checks
  • Email alerts
  • 5 status pages
Start free

Most popular

Pro

£29/month

  • 200 monitors
  • 30-sec checks
  • All 16+ alert channels
  • 90-day data retention
Get started

Enterprise

£89/month

  • 500+ monitors
  • 30-sec checks
  • SSO & audit logs
  • Dedicated support
Contact us

All plans include multi-location checks, incident management, and public status pages. Full plan comparison →

Frequently asked questions

Everything you need to know about multi-location monitoring.

Where are your users having a bad time?

Monitor from 10+ locations worldwide. Free for 50 monitors, no credit card required.

Start global monitoring free