This post is an English summary of my presentation on capacity management, prepared for an IT Service Management course based on the ITIL framework. The full slide notes are available in the Turkish version.

What is capacity management?

Capacity management is an ITIL process that runs through the entire service lifecycle. Its job is simple to state and hard to do well: make sure that IT services and the infrastructure behind them always have enough capacity and performance to meet agreed business needs, on time and at a justifiable cost. If a service cannot deliver the capacity and performance it promises, the business simply does not get the value it expects from it.

Goals

  • Produce and maintain an up-to-date capacity plan that reflects the current and future needs of the business.
  • Act as the go-to point for advice on all capacity and performance questions, for both the business and IT.
  • Make sure service performance meets all agreed targets by managing the capacity of both services and the resources behind them.
  • Help diagnose and resolve capacity- and performance-related incidents and problems.
  • Assess the impact of every change on the capacity plan and on service performance.
  • Use proactive measures to improve performance where it is cost-effective to do so.

Scope

Capacity management looks at every resource needed to deliver an IT service, in the short, medium and long term: hardware, software, networks, data, applications and even people, since staffing and skill levels are also a form of capacity. In practice, it covers monitoring usage and workload patterns, tuning existing resources, forecasting future demand, influencing demand together with financial management, producing a capacity plan aligned with SLAs (service level agreements) and SLRs (service level requirements), and keeping an eye on new technologies that could deliver services better or more cheaply.

Value to the business

  • Fewer capacity- and performance-related incidents, and therefore better availability.
  • The required capacity and performance, delivered at the lowest reasonable cost.
  • Higher customer satisfaction and user productivity, because service levels are met.
  • Smoother introduction of new or changed services, thanks to early involvement in design.
  • More reliable IT budgeting, based on a forward-looking capacity plan.
  • The ability to follow a greener, more responsible IT strategy.

Key concepts: it is a balancing act

Capacity management constantly balances two things: cost against capacity (every unit of capacity you buy should be justified by a business need and used efficiently) and supply against demand (the IT processing power available today and tomorrow should match what the business actually needs, and sometimes demand itself has to be managed). Managing capacity in large, distributed environments is complex, which is exactly why growth should be planned rather than reacted to. Good capacity data answers practical questions: which components need upgrading (memory, storage, CPU, bandwidth), when, and at what cost. Without it, other processes such as change management, service level management, problem management and IT service continuity all become less effective.

The three sub-processes

  • Business capacity management translates business plans and needs into service and infrastructure requirements, so that future demand is understood, planned and delivered in time.
  • Service capacity management manages, controls and predicts the end-to-end performance of live IT services against the targets in SLAs and SLRs.
  • Component capacity management manages, controls and predicts the performance, utilization and capacity of individual technology components, ideally with automated thresholds and alerts.

Capacity management sub-processes

The tools used for capacity management should fit the organization’s management architecture and integrate with other monitoring and automation tools. Monitoring and control activities in service operation, along with teams such as network and server management, provide much of the data the process relies on.

Proactive and reactive activities

Some capacity management activities are reactive, dealing with issues as they appear; others are proactive, preventing them in the first place.

Proactive: spotting performance issues before they happen, trending current usage to forecast future needs, modeling the impact of planned changes (including retiring services), budgeting and implementing upgrades before SLAs are breached, and continuously looking for cost-effective ways to improve performance.

Proactive capacity management activities

Reactive: monitoring, measuring, reporting on and reviewing current performance, responding to threshold events with corrective action, and helping resolve specific performance problems escalated from the service desk.

The more successful the proactive and predictive work is, the less reactive work will be needed.

Reactive capacity management activities

Ongoing activities and monitoring

The biggest challenge is understanding the link between business demand and IT workload, and translating it into the impact on services and resources. Typical metrics include CPU and memory utilization, I/O rates and device utilization, queue lengths, disk usage, transaction rates, response times, batch durations, database and directory usage, concurrent users and network traffic.

Ongoing capacity management activities

Comparison with my own work experience

Business capacity management: We used various measurement systems to track our work, but the general problem I saw was that the data collected received little attention and was rarely read. Weekly IT help desk reports on my daily work were emailed to our managers, but there was little interest in performance evaluations based on this data, in corrective actions for the most frequently reported issues, or in preventive measures.

Monitoring: We used separate systems to monitor hardware devices, databases, network performance and external attacks, which made monitoring harder and more cumbersome. There was no budget for a system that would let me monitor everything from a single portal, so active tracking and instant alerting were limited.

Service management: Our servers’ disaster recovery setup existed only within the company’s own premises. Due to budget constraints, there was no off-site disaster recovery. This should be put on the agenda as soon as possible, and a disaster recovery site in a different geographical location must be set up.

Service desk: All of our users’ requests arrived marked as urgent. Users opened tickets or picked up the phone without trying any basic fixes first. Training users on what they can do themselves when they run into common problems would allow many issues to be solved before a ticket is even opened, reducing the service desk’s workload.

bilgihizmetkapasiteteknolojilerindeyonetimi