EnterraHost
A domain is checked from one region by default. Adding a second means the same site is tested from two places on the network, and the two answers do not always agree.
Open the domain, then the monitoring region field, and choose a different region from the list.
The five regions
| Region code | Location |
|---|---|
AF-CPT | Cape Town, South Africa |
US-STL | St. Louis, United States |
EU-FRA | Frankfurt, Germany |
OC-SYD | Sydney, Australia |
AS-SIN | Singapore |
Cape Town is the default. The choice is about the audience, not about you.
What a second region actually buys
It separates two failures that look the same from one location.
A problem on the path. If a site fails from Cape Town and answers from Frankfurt, the server is up and something between Cape Town and the server is not. That is a routing or transit fault, it is usually temporary, and it is invisible with a single region because all you see is a down alert.
A problem at the server. If both regions fail, the site itself is not answering. That is the case where the alert is about your code, your certificate or your host.
The distinction matters most when you are trying to get somebody else to act. “Our monitor in Frankfurt cannot reach it” is a support ticket with a direction in it. “It was down for four minutes” is not.
What it costs
Two regions mean the same site is checked twice as often, against the plan’s check allowance, and it produces two sets of latency figures rather than one.
The domain count does not change. A second region is a property of the same monitored site, not a second site, so it does not consume another slot.
Reading two regions against each other
The useful comparison is the gap, not either number on its own.
Two regions a continent apart will not show the same latency, and that is expected. What is worth acting on is a gap that opens suddenly, or one that closes. A site that normally answers in 40ms from Frankfurt and 180ms from Sydney, and is now 180ms from both, has moved rather than improved.
Where the site is behind a CDN, both regions may answer from a nearby edge and the latency tells you about the edge rather than the origin. The origin question is a different one, and does a CDN fix TTFB is the article that deals with it.
When one region is enough
If your audience is in one country and your hosting is in that country, a second region adds cost and noise. You will get alerts about distant routing that your visitors never experience.
Start with one. Add a second when you have a reason, which in practice means an audience in two places, a CDN you want to check from more than one edge, or a hosting provider whose network you have started to distrust.
Latency figures from a single region still repay watching over time, because the interesting signal is usually the change rather than the absolute number. Reading a latency alert covers telling a path problem from a slow server, and understanding the report covers how the timing numbers are produced.