Infrastructure · monitoring · escalation
Servers watched around the clock
Automated checks on every server and on each service running on it, an alert the moment one fails, and an escalation path if it is not cleared.
Checking every service on a single server 24 hours a day is difficult. Doing it by hand across a fleet of servers is impossible. An effective monitoring system is what keeps uptime from depending on somebody noticing.
A company without a good monitoring system — or worse, without one at all — has longer downtimes and carries more risk from service disruptions. An undetected minor issue turns into a major one quickly, and the damage grows with it.
What we run
Our infrastructure is built from high-availability clusters of different machines, running varied operating systems and applications.
Our monitoring systems and tools give our system administrators one view of the health of that infrastructure. We watch a large number of parameters on each server and on each of the individual services that sit on it.
Services monitored
What is checked, on every server
Server resources
- Server disk space
- Server CPU usage
- Server memory usage
Network and DNS
- Network connectivity
- DNS services
Web services
- HTTP
- HTTPS
- FTP
Mail and databases
- SMTP
- POP
- IMAP
- MySQL
- MSSQL
When something fails
How an alert escalates
An open issue never stays with one person.
An alert goes up at once
If a server or a service fails, or a resource crosses its set limit, a notification appears immediately on the screen of every member of our infrastructure monitoring staff.
It escalates if it is not cleared
The notification system handles escalation too. An issue still open after a set interval sends SMS alerts to higher-level system administrators.
Then it reaches management
If it is still open after that, the alerts carry on up to management, so no open issue stays with one person.