When you have a web page and/or application you want to be available each second, 7 days a week.
Our NOC have a medium availability of 99,9%, this means that are months when the availability it is 100%, other moths less.
Why all this, and who tell those numbers.
It is really important to us to know that everything go fine. The official numbers are taken using a third party provider: Pingdom. This because the data should be just right. They could tell us and our customers the situation from outside of our network.
On the other side we are monitoring our network, each server or network device. There it is a lot of data monitored, but as you could guess we are interested only on yellow and red tickets from our Data Center. When something become yellow means that something could go wrong, this could be a damaged harddisk, a damaged network cable or optical fiber, it could also be some very high traffic, or anything else. But we have to solve very fast the ticket to not become a red one.
When a ticket goes red it means that an important problem appeared on the NOC and that this problem should be solved in minutes. If the problem it is too big we are going on backups in order to have no downtime.
Another important fact in our NOC it is the traffic. We monitor the traffic to keep it under 50% load. When something it is at 50% load it means that we should think to an upgrade. When the medium load it is at 75% it means that we are making the upgrade.
Do you think that that’s all?
No. We have another issue to check ….. the response time of our servers. We are working to keep it under 1 ms. But this also depends on the programing of the webpages. So where we see that the response time it is more than 1 ms we start to find solutions where are possible. Even the clients don’t want to remove any lemon code from the pages we try to find another solutions.
Because of these each time a client move the webpages on our servers will see a faster response time. At our data center the web page could have a 0.20s response time which means an answer of about 200ms from our servers (it could vary in normal time between 89-200ms) as you could see direct from our third party monitor https://uptimebutler.com/checkpublic/7cde5a9be3ee70a96e14899a On the same page you’ll see our uptime from outside of our datacenter.
Not at last, one of the most important are the DNS servers. We are using at the moment more than 15 name servers to have your domain changes in minutes, not in 24 hours. The uptime of DNS it is guaranteed at 1000%!