Multi-region website monitoring that confirms outages

Before OctoPing tells you a site is down, it compares the failure with what the US, Europe and Asia see, so you can say whether the site is down for everyone or only for one network.

When one checker says down and visitors say fine

A monitor that checks from a single place reports whatever that place sees. A firewall or bot rule can refuse that one checker's address and return a 403 for hours while visitors browse the site normally. A CDN or routing fault can make a site unreachable from Europe while Asia loads it. In both cases a single-location monitor sends a down alert, and you apologise to a client for an outage their customers didn't experience.

Four vantage points

Each site is checked from four vantage points. All four send the same request, with the same headers, timeouts and retry policy, so a difference between them describes the site and not the monitor.

The primary checker runs at the interval you set for the site and drives alerting. The regional probes check on their own 5-minute rhythm, and what they report is used to confirm or contradict a failure the primary has seen. A region that fails while the primary still reaches the site doesn't send a down alert on its own, and its results still show on the site's page.

How a failure is confirmed

The primary checker decides when to look closer. When it sees a failure, OctoPing weighs it against the reports the regions have sent since the failure began.

• A region also fails. The failure is confirmed and the normal down alert goes out. The alert lists each location with its result. When some locations still reach the site, the subject names the ones that don't, and the email says the cause may be a regional network, CDN or DNS problem.
• Every region reaches the site. The failure belongs to the primary checker, so there's no down alert, no recovery alert and no downtime counted. The site shows a Partial status, in amber, and you get one notice per run saying the checker is being refused. The notice names the HTTP status it got and the regions that see the site up, and it lists the addresses to allow in your firewall. If a region starts failing later, the normal down alert follows.
• No region has reported yet. The alert waits for reports for up to 10 minutes and then goes out as usual. A site with no regional data alerts normally, and a region that has gone quiet stops counting as evidence, so a broken probe can't hide an outage.

Once an outage has been declared, later healthy reports from a region don't undo it. Regions decide whether a failure counts as an outage, and the first failing check and the first recovery still mark when it started and ended.

What you see

The site's page lists each location as Up or Down, with the HTTP status or failure reason and how long ago it checked. When the status is Partial, the page explains that the primary checker is being refused while every region sees the site up. Public status pages show a Partial site as operational, because visitors can reach it.

How it fits the price

All four vantage points check every site on every plan. Plans start at $19 a month for 20 sites. See pricing for the full list.

Every account starts with a 14-day free trial. No card required. Start your 14-day trial and add a client's site to see it working.