Overview of APM Admin Philosophy

Document ID : KB000017359
Last Modified Date : 14/02/2018
Show Technical Document Details

Description:

This is an expansion of a Tuesday Tip sent to the Community Site

Solution:

Operating System Administration and APM Administration use similar approaches. For APM Administration, it comes down to making a choice:

Will I do the MINIMUM needed and then "pay as you go" when things go wrong?

Will I have a REACTIVE or PROACTIVE APM environment?

This is summarized in the following table:


Philosophy Admin Activities System Stability Use of Tool Time-Consuming
Minimum/ Do Nothing Only if absolutely necessary. System runs until it breaks. Not optimized. Generate metrics Only if fixing an APM system.
Reactive Fix broken components. Add agents, dashboards etc. without being concerned about implications Visible issues fixed but deeper ones may not be addressed. Generate metrics Minimum reporting Only if fixing an APM system or resolving upgrade issues. In "Reactive Hell" state because not applying lessons learned during outages.
Proactive Frequent, to have optimized and current systems. Good because visible and deeper issues addressed, architecture and configuration is optimized Generate metrics Full use of reporting Event correlation Application optimization Capacity planning Takes full advantage of the solution. Time spent on proactive tasks is offset by reduced outage time.

References

https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/4/html/Introduction_To_System_Administration/ch-philosophy.html -- Red Hat Document on System Administration Philosophy

http://kagan.mactane.org/essays/sysadmin.php -- Another Overview on OS System Administration