Compile Uptime Notifications into One Message / Monitor dependencies / Cascading monitoring#129
It would be great if instead of spamming our slack channel with hundreds of messages when a server goes down it would instead just group all uptime notifications with the same ip into a single message and say something like “Multiple sites on the ip 127.0.0.1 are down:” and then list all the sites that are down inside the message. This way your slack won’t crash along with the server from overloaded notifications.
Agreed, but I think this needs to be explicitly set to a group not determined by public IP but rather dependent checks. Simple case, if router fails and causes multiple checks to fail, we currently get that many alerts. Having a critical dependent check of pinging the router for example will supress the notification about 10 sites out and simply send an alert for the router failure.
This is a good idea please could you add it to the roadmap?
The ability to group different uptime checks into a single alert is a blocker for my company to adopt uptimerobot.
For SSL/Domain expiration notifications please group by the domain/cert and do just one. I got 377 pagerduty alerts last night for a single domain that expires in 30 days
I have two devices at the edge… if either one fails, it causes every monitor to go down and every possible alert to get sent. While I realize in Nagios, its based on timing (check child before parent causes an un-due alert) it would be nice to be able to link monitors in a parent/child relationship, or even a dependency of some sort.
Another vote here. In a basic example, we’re pinging a WAN IP address, but we’re also monitoring a service (ie, HTTPS) on that IP address. On its own, if the service was to go down, we would get a single alert, but if the WAN IP address, we get 2 alerts for ultimately the parent going down. I think if we could add dependancies or a parent/child relationship to prevent excessive and unnecessary alerts/notifications, that is key.
Hey everyone.
We’re exploring options for this feature and would like to get some more input from you. If you’re up for a quick chat, grab some time with me here: https://calendar.app.google/yH2L57oMpt6E5AxEA
Or if a short survey is easier:
https://info450823.typeform.com/to/UqfApPf4
Either way, really appreciate your input!
@Tutku That would be very usefull feature. One of the most crucial for me. For example in my case, I have multiple monitors for different URLs within same website. Sometimes only specific pages has outage. And now if whole website is down, I’m receiving many alerts which is annoying and spending my phone calls budget.
I also have multiple websites per client, which are hosted at the same hosting. And if whole hosting has outage, I’m reciving 10 alerts per 1 client.
Hello, what’s the status of this idea?
Thanks everyone for the feedback and discussion on this request.
We’ve recently released Alert Storm Protection, which reduces alert fatigue by automatically suppressing notification floods when many monitors fail within a short period. While it doesn’t introduce monitor dependencies, it does address many of the situations where a single infrastructure issue would otherwise generate dozens of alerts.
You can read more about how it works here:
https://help.uptimerobot.com/en/articles/15433157-alert-storm-protection-in-uptimerobot
We’ll keep this request open because many of you are specifically asking for dependency-aware monitoring (for example, suppressing child monitor alerts when a parent monitor is down). We’d love to hear whether Alert Storm Protection covers your use case, or if explicit monitor dependencies are still something you need. If the latter, sharing your setup or workflow would help us understand which scenarios aren’t covered today.
That Alert Storm Protection will definetely help with certain cases yes. But my specific use case for Uptime Robot actually would need both kinds of solutions. A Storm Protection, as well as a dependency option for if the main router is the one that goes down.