Uptime Robot has been a very stable service for a long time. Yet, for the last few weeks, we know that some users have received false-positives.
Why did that happen?
In short, the servers we used couldn’t handle the load :).
In detail, we always look for ways to heavily optimize the system and the engine to minimize the need for new CPU, RAM or diskspace. The optimizations vary from perfecting the queries used or compressing not-frequently-used-data to choosing the right platforms for the right job. But, we experienced that, how much you optimize, every system has a limit and you shall grow that system before getting close to that limit.
What are the actions taken?
We have added more servers and strengthened the current ones.
And, besides the servers added, from now on, we’ll be adding new ones before the load gets heavier so that the system will always stay healthy.
Additionally, various new controls are added to the engine to minimize false-positives and Uptime Robot will “very” soon start verifying downtimes from multiple locations at once (rather than the current setup which is 1 remote location).
What are the new IPs (if you need to whitelist them)?
There is now an additional engine and 4 more remote nodes to verify downtimes. The up-to-date list of IPs can be found at Locations & IPs page.
To sum up, just wanted to inform you that our hands are on it.
