Operational Excellence - The Practice of Cloud System Administration - page 415

Tobegin,theteamshouldenumeratethemajorservicesitprovides.Forexample,ateam
that is responsible for a large web site might determine that each web property is a service,
each internal API is a service, and the common platform used to provide the services is a
service. There may be multiple platforms, in which case each is considered a service. Fin-
ally, the infrastructure itself is often counted as a service as far as assessment is concerned.
Assessmentsshouldbedoneonaperiodicandrepeatableschedule.ThefirstMondayof
each month is a common choice for frequency. The team meets and conducts a self-assess-
ment of each service. Management's role is to maintain high standards and to ensure con-
sistency across the services and teams. Management may set global standards for how cer-
tainORsareevaluated.Forexample,theremaybeacorporatechangemanagementpolicy;
compliancewiththatpolicyshouldbeevaluatedthesamewayforallservicesbyallteams.
Eight core ORs should be used to assess all services. Appendix A includes details about
these eight operational responsibilities, along with questions to ask during assessment to
aid the team's understanding of the OR. The questions are followed by look-for's describ-
ing behaviors typically seen at various levels. Look-for's are indicators that the service is
operating at a particular level. They are not checklists of behaviors to emulate. Do not at-
tempt to achieve every look-for. They are indicators, not requirements or checklists. Not
every look-for is appropriate for every service.
In addition to the eight core ORs, Appendix A lists other ORs that are particular to spe-
cific services. They are optional and serve as examples of service-specific ORs that an or-
ganization may choose to add to the core eight. Organizations may also choose to invent
ORs for their own special needs.
As the monthly assessments progress, the changes over time should be apparent. The
results of the assessments will help teams determine project priorities. Over time, the roll-
ups described in this chapter can be used to compare services or teams.
20.8 Summary
In this chapter we discussed how to assess the quality of operations and how to use this
assessment to drive improvements.
Measuring the quality of system administration is complex. Therefore for each service
we assess eight different qualities, called operational responsibilities: regular tasks (how
normal, nonemergency tasks are performed), emergency response (how outages and other
emergenciesarehandled),monitoringandmetrics(collectingdatausedtomakedecisions),
capacity planning (determining future resource needs), change management (how services
are purposefully changed from birth to end), new service introduction and removal (how
new products, hardware, or services are introduced into the environment and how they are
removed), service deployment and decommissioning (how instances of an existing service