Child pages
  • Operation and Hosting Packages - Monitoring Concept
Skip to end of metadata
Go to start of metadata

Monitoring your application

Critical components and key metrics of our hosting and operation packages for your Atlassian applications are monitored around the clock by our central monitoring system. Depending on your applicable SLA, our system administrators are notified of critical incidents by SMS within service hours and immediately start working on them.

Systems that we operate as virtual appliances in your customer infrastructure as part of our operating package are also automatically monitored via a Site2Site VPN connection.

The following components and metrics of a hosting or operation package are monitored by our monitoring system:

Category

Type

Metric

Web server

Application availability

Can the login page of your application (e.g. Confluence) be opened within 10 seconds?


Simultaneous connectionsDoes the number of simultaneous connections reach critical limits?

SSL certificate expiry dateIs the SSL certificate used for secure transmission still valid for an adequate period of time?
AtlassianJava heap memory Is there still enough memory available for the application in the Java virtual machine?

Threads, open filesAre the number of open files and running threads (processes) reaching critical limits?
Data securityLocal ZFS snapshots Were file system snapshots taken at the scheduled intervals?

Mirroring to backup server Was a data backup to one of the offsite backup servers performed within the previous 24 hours?
SystemCPU utilizationIs the CPU utilization reaching critical limits?

Disk spaceIs there still enough disk space available for your application?

Internet access Can the server system access the public internet?

Server timeAre the services for setting the server time running correctly?

Email dispatchAre the local email services on the server operational?

In addition, metrics from the above categories which take up to 60 seconds to query are regularly checked and logged to save time when troubleshooting.

Below is a sample of the monitored data (this will vary depending on the system):

  • No labels