SLM can be considered QoS monitoring and management based on key performance indicators (KPIs). QoS KPIs can range from generic availability and usage statistics to entity-centric per-interaction indictors. Adequate SLM requires potential problems identification — such as gradual performance degradation — and alerts creation enabling downtime risk minimization. Consequently, SLM practices should include comparing actual performance to pre-defined expectations, determining appropriate actions and generating expressive reports to permit service improvement.
“View Part I of the Service Level Management of Cloud Computing series here“