Multi-region monitoring

Every check runsfrom every region

Plenty of monitoring services list a dozen locations and check from one at a time, or call the others in only after a failure. Sentinel runs every scheduled check from all fifteen regions on six continents, spread out so they never land on your server as one burst. You get fifteen views of every check, and an outage is confirmed before anyone gets paged.

Sentinel's network

15 regions

AshburnChicagoSeattleTorontoMexico CitySão PauloLondonFrankfurtTel AvivJohannesburgMumbaiSingaporeTokyoSeoulSydney
15 regions 6 continentsEvery region, every cycle

How it works

Global checks without global false alarms

Running every check from fifteen regions gives you more evidence. It shouldn't give you fifteen times the alerts. Sentinel marks a monitor down only when at least half of its regions agree it's unreachable, with three or more of them on two different continents. With all fifteen regions on, that's eight.

A failure in one region stays visible in that region's results on the monitor, but it doesn't page anyone. When Frankfurt can't reach your site and the other fourteen regions can, you're looking at a regional problem, not an outage.

And when one region is the business, say most of your customers are in Europe, add a second monitor scoped to just that region. It alerts on that region alone, with no consensus required. One monitor for the global truth, another for the region you can't afford to lose. Each alert means exactly what it says.

  • Every check runs from every region
  • At least half the regions, on two continents, confirm an outage
  • One region's failure stays visible without paging anyone
Exactly how the regions confirm an outage →

Alert decision

Regions checked
15
Failing
1 (Frankfurt)
Reaching the site
14
Needed for an outage
8, on two continents
Verdict
Regional, no alert

Fourteen regions still reach the site, so nobody is paged. Frankfurt's timeouts stay on the monitor for you to see.

Features

Reliable alerting through global verification

Fifteen regions, six continents

Ashburn, Chicago, Seattle, Toronto, Mexico City, São Paulo, London, Frankfurt, Tel Aviv, Johannesburg, Mumbai, Singapore, Tokyo, Seoul and Sydney. Fifteen independent views of every single check.

Fewer false alarms

A site is only marked down once at least half its regions, on two continents, confirm it. A network blip in one data center won't page your on-call team at 3am.

Per-region visibility

Every region's latest result sits side by side on the monitor view. If your site times out in Europe but responds in North America, you'll see exactly that at a glance, while alerts stay reserved for outages the regions confirm together.

CDN verification

Verify your CDN serves content correctly from every edge location. Catch regional CDN misconfigurations, expired edge caches, and routing issues before your users hit them.

No rotation

Multi-region shouldn't mean taking turns

Having locations all over the map doesn't mean they're all checking your site. Some services rotate scheduled checks between regions. Others check from one location and call the rest in only after it reports a failure. Sentinel runs every scheduled check from every region.

SentinelRotating locationsConfirm on failure
Regions on each scheduled checkAll fifteenOneOne
The other regionsCheck every intervalWait for their turnOnly called in after a failure
A problem only Frankfurt sees shows upWithin one intervalWhen Frankfurt's turn comesOnly if the main location fails too

On a 30-second interval, all fifteen regions check your site every 30 seconds. A monitoring system shouldn't become a load test, so the regions are spread out within each interval. Your server never gets fifteen requests at the same instant, and no region ever skips a turn. A timeout in Frankfurt or a DNS failure in Singapore shows up on the next check, not whenever that region's turn comes around.

Global infrastructure

Fifteen regions, six continents

Every location takes part in every scheduled check.

North America

  • Ashburn
  • Chicago
  • Seattle
  • Toronto
  • Mexico City

South America

  • São Paulo

Europe

  • London
  • Frankfurt

Middle East

  • Tel Aviv

Africa

  • Johannesburg

Asia

  • Mumbai
  • Singapore
  • Tokyo
  • Seoul

Oceania

  • Sydney

15

Monitoring regions

6

Continents

15

Regions on every check

FAQ

Multi-region questions.

Still have questions? Get in touch.

Why does Sentinel monitor from multiple regions?

One location only gives you one view of the internet. A network blip or routing issue near that location looks exactly like your site going down. Checking from fifteen regions shows you which it is, and an outage has to be confirmed across regions before anyone is alerted.

How many regions does Sentinel check from?

Fifteen: Ashburn, Chicago, Seattle, Toronto, Mexico City, São Paulo, London, Frankfurt, Tel Aviv, Johannesburg, Mumbai, Singapore, Tokyo, Seoul and Sydney. Every check runs from all fifteen, which is the number that matters. Many monitoring services list a dozen locations and check from one of them at a time.

Does Sentinel rotate monitoring locations?

No. Sentinel doesn't rotate scheduled checks between regions. Every scheduled check runs from all fifteen regions, so on a 30-second interval every region checks your site every 30 seconds. The regions are spread out within each interval, so your server never gets fifteen requests at the same instant, but no region ever skips a turn.

Are the other regions only used to confirm failures?

No. All fifteen regions take part in normal scheduled monitoring, not just in confirming a failure. Consensus decides whether a failure becomes an outage alert, but every region reports on every check either way.

How does consensus prevent false alarms?

A monitor is marked down only when at least half of its regions agree it's unreachable, and at least three of them are on two different continents. With all fifteen regions on, eight have to agree. A transient issue in one region stays visible on the monitor but won't wake your on-call team.

Will Cloudflare or a WAF mistake the checks for bot traffic?

SentinelBot is a verified bot in Cloudflare's Radar directory, so checks from all fifteen regions are recognized as legitimate monitoring traffic on Cloudflare-protected sites instead of being challenged. The user agent is documented and Sentinel's monitoring IPs are published for allowlisting on other firewalls.

Can Sentinel tell me if an outage is regional?

You can see it. Because every check runs from fifteen regions, the monitor view shows each region's latest result side by side, so a site that's down in Europe but up in North America looks exactly like that. Alerts are reserved for outages the regions confirm together, so a blip visible from only one region won't page you. If a specific region matters enough to alert on, such as the market your customers are in, add a second monitor scoped to just that region and it will alert on that region alone.

How often do multi-region checks run?

The check interval depends on your plan, from every 5 minutes on the free plan to every 15 seconds on Business. Whatever the interval, every region checks at it. More frequent checks mean faster detection and faster confirmation of real outages.

What happens after a real outage is confirmed?

Once an outage is verified across regions, Sentinel sends alerts by email, SMS, Slack, Discord, Telegram, or webhook and opens an incident with an automatic timeline of which regions were affected.

Monitor from everywhere your users are.

Start monitoring from fifteen global regions today: reduce false alarms, detect regional outages, and verify CDN performance worldwide.

No card for the free plan · 14-day trial on paid plans · Cancel anytime