Skip to main content

System Status Graphs

The system keeps track of critical resources over a period of 24 hours and shows those as graphs to system and tenant administrators. The graphs are available on system and tenant level; some graphs are also available on extension and trunk level. They are shown in the web interface. The system level graphcs can be sent to the system administrators at midnight by email.

The following graphs are available:

  • Calls: The calls graph shows how many calls were made. There are two graphs. The first one shows how many call legs were active and the second one shows how many calls were in use. The call leg count is important because it is better resource indicator than the number of calls.

  • Call Quality: The system estimates the call quality on a MOS scale. This makes it easy to spot quality problems on system, tenant and also extension and trunk level.

  • Registrations: The system keeps track on registrations and subscriptions. Ideally this graph shows a flat line, however in reality as devices sign in and out there are changes in the graph. Subscriptions include BLF and MWI subscriptions.

  • Registration Changes: In order to amplify registration changes there is a seperate graph that shows the changes in registrations. Even though the total number of registrations may be unchanged, when one devices registration expires and another device signs on, this graph will show the changes.

  • Media CPU Usage: This graph shows the usage of the RTP processing thread of the PBX in terms of the ratio between active and total time. This makes it easier to see when the system is reaching maximum call capacity and will eventuall start to reject calls.

  • Main CPU Usage: This graph shows the usage of the main thread of the PBX. Like with the media CPU usage, this shows the percentage of idling and active CPU usage. Because the main thread runs with a lower priority than the media thread, the peak usage is typically significantly higher than the average usage. Average usage of more than half the CPU would raise a concern.

  • Memory: The memory graph shows how much memory the system process has allocated. Because of the way the system allocates memory, the total amount is usually not shrinking; this is not a memory leak. However the total amount of allocated memory should after a few days stabilize at a certain point.

  • Faults: The system keeps track of page faults and delays in processing. Page faults happen when the operating system is trying to save memory and transfers parts of the PBX process to the file system. It is okay if this happens occasionally; however if there are many page faults the system should have more memory. Processing delays indicate situations when the internal event loop takes longer than one second. This should happen only in rare events such as software updates and restarts. Otherwise the server performance needs to be reviewed.

  • HTTP/S: The graph shows how many HTTP and HTTPS packets were processed by the system.

  • SIP: The graph shows how many SIP packets were processed by the system on UDP, TCP and TLS level.

  • T/FTP: The graph shows how many FTP and TFTP packets were processed by the system, if at all.

  • LDAP/S: The graph shows how many LDAP and LDAPS packets were processed by the system.

  • CSTA & TCP-SYN: This graph shows how many CSTA packet were processed and how many times the system accepted TCP connections.

The graphs are also available through the REST API and can be displayed on external tools like Grafana.

Last updated on