Skip to content
Technological controls
ISO 27001A.8.6

ISO 27001 Annex A 8.6 Capacity management

What it says

The use of resources shall be monitored and adjusted in line with current and expected capacity requirements.

Quoted as published. Everything else on this page is written by Control Stack.

In plain English

Monitor the use of resources, including systems, people and facilities, and adjust it in line with current and expected capacity requirements.

Your organisation must monitor the use of resources and adjust it in line with capacity requirements, meaning how much of each resource is needed, both now and as expected in future. Adjusting can mean adding capacity or reducing demand. Resources go beyond computing:

  • Systems: processing power, memory, storage and bandwidth.
  • People: the numbers and skills that services depend on.
  • Facilities: offices and other physical space, such as server rooms.

Why it matters

A full disk, an exhausted licence pool or the loss of the one person who runs a system can stop a service with little warning. Watching use against expected demand can show the limit early enough to act, including on items slow to buy.

How to implement

  1. Set capacity requirements, starting with the services the business relies on most.

    For each service, record the systems, licences, people and space it depends on, and the level needed at normal and peak times. Your organisation sets these levels, because the control sets none. Give the most critical services the most detail.

  2. Monitor use against those requirements, with alerts.

    Use a monitoring tool for processor, memory, storage and bandwidth, with alert thresholds that leave time to act. For a bought service, use the provider's usage reports and check its quotas and plan limits. Keep a list of one-person dependencies, and track workload and floor space through ticket queues, rosters and floor plans.

  3. Forecast expected demand before it arrives.

    Combine usage trends with known changes, such as new systems, projects, customers or sites. Flag anything slow or expensive to obtain, such as hardware or hard-to-find skills, so the decision is made early enough.

  4. Adjust by adding capacity or reducing demand.

    Add processing, storage, a higher service tier, cloud auto-scaling, people or space. Or reduce demand: delete obsolete data, retire unused systems, move batch jobs, tune queries or limit non-critical bandwidth use. Send system changes through change management (Annex A 8.32).

  5. Test that peak demand can be met.

    Run load or stress tests on systems you operate, in a planned window. For a bought service, get the provider's permission first, or compare its published limits and service levels with your peak.

Keep it running

Revisit forecasts and thresholds when a new system, project or customer changes demand, or when an alert shows the estimate was wrong. Update the list of one-person dependencies as people change roles.

Evidence and audit

  1. Check 1 of 5What an auditor asks for

    Ask for

    Capacity requirements, compared with the list of business services.

    Look at

    Whether each service on the list has a stated level, with the most detail for critical ones, and whether people and facilities are covered as well as computing.

    Good evidence

    Each service on the list has a recorded requirement covering the systems, people and space it depends on, with critical services in the most detail.

  2. Check 2 of 5What an auditor asks for

    Ask for

    Live monitoring dashboards and alert settings for a sample of services.

    Look at

    Whether each resource in the requirements is watched, including provider usage reports for bought services, and who receives the alerts.

    Good evidence

    Each required resource is monitored against its requirement, and recent alerts reached someone who acted on them.

  3. Check 3 of 5What an auditor asks for

    Ask for

    The latest capacity forecast, compared with the project list and planned changes.

    Look at

    Whether known changes are counted, and whether items slow or expensive to obtain are flagged early enough.

    Good evidence

    The forecast shows expected requirements, reflects each sampled planned change and leaves time to act on slow items.

  4. Check 4 of 5What an auditor asks for

    Ask for

    Adjustments made after recent alerts or forecasts, such as upgrades, clean-ups or rescheduled jobs.

    Look at

    Whether each alert or forecast led to a change or a decision before the limit was reached.

    Good evidence

    Each sampled alert or forecast has an adjustment made before the service was affected, or a recorded decision not to act.

  5. Check 5 of 5What an auditor asks for

    Ask for

    Results of the last peak-load test, monitoring data from the last real peak or, for a bought service, the provider's published limits.

    Look at

    The tested or observed load or published limit, compared with the peak in the requirements.

    Good evidence

    The evidence shows the peak requirement can be met, and any shortfall has an action with an owner.

Mappings

Closest match in ASD Information Security Manual: ISM-2091, Enforce resource limits for AI models, which this control covers in full and more.

ISO 42001

ISO/IEC 42001: 1 mapping

  • A.6.2.6AI system operation and monitoringPartially overlaps
ASD ISM

ASD Information Security Manual: 6 mappings

Mappings show how controls relate across frameworks. They do not imply full equivalence or certification.

Get help with this control

ISO 27001 training

Mindset Cyber, a PECB Authorised Titanium Partner, sells official PECB ISO/IEC 27001 courses that map directly to the controls in this library.