Capacity Planning for your Sitecore: Performance vs. Cost

October 1, 2019

Capacity is defined by your infrastructure’s ability to handle the expected traffic based on a given configuration.


Microsoft Azure allows web applications like Sitecore - when in PaaS configuration - to be “right-sized” for the task at hand. This means it will serve web pages, CMS and/or Marketing applications to users. Scaling those independently during a given day, hour or minute is of huge benefit to any organisation.

When we engage with customers, the group will define the parameters that we wish the environment to adhere to, automate where possible and then review these parameters based on environment performance on a regular (quarterly) basis.

Here, we’re breaking down the cost-led, performance-led and blended approaches an organisation can take when deciding on their capacity requirements.

Cost Lead
Performance Lead
Blended
Driver
Cost Lead

Cost must hit a fixed budget regardless of customer experience requirements of the platform.

Performance Lead

Environment must perform at highest level, always, regardless of cost.

Blended

Value

Approach
Cost Lead

Resources configured with bare minimum sizing, does not scale manually or automatically.

Performance Lead

Maximise size of all resources, do not implement auto or manual scaling.

Blended

Right-size resources to cope with predictable averages.

Implement auto-scaling with cost caps to cope with peak demand.

Setup alerts when environment is at capacity, enable overrides.

Impact
Cost Lead

Application is slow, barely functions is often.

Performance Lead

Cost is significant. Under used capacity is paid for.

Blended

TCO is reduced, UX is consistent.

Mindset
Cost Lead

Cost-led organisations, IaaS mindset, CapEx mindset.

Performance Lead

Suffered pain before. Not budget conscious.

Blended

OpEx rather than CapEx mindset. Able to agree on predictable but changeable costs.

Our recommendation

The Blended approach is highly recommended by Dataweavers for any client wishing to deploy on Azure PaaS. It enables the organisation to serve fast, low latency web pages and consistent user experiences when under peak demand, without having to incur the costs of permanently scaling secondary roles, such as analytics processing.

Sitecore 9 on Azure PaaS does provide almost infinite control over the way Sitecore works, however the fundamental basics of Sitecore scaling have not changed and there are some common concerns that we see when considering capacity planning.

Here’s what we think all organisations should be considering:

Primary concerns
  • Web Applications
    • Content Management Server response time
    • Content Delivery Server load times
    • Analytics reports up to date
  • Azure Search
    • Content Indexes up to date
  • SQL Server
    • Content Databases can publish quickly
    • xDB Analytics growth
Secondary concerns
  • Web Applications
    • xDB Marketing Automation
    • Email dispatch
  • Azure Search
    • Indexing speed during publishing
    • Indexing speed during peak analytics load

The trick is to balance cost of infrastructure with both peak demand and quality of service. The environment will be scaled initially to meet the load times as defined in the business requirements, consistent with the traffic expectations agreed during the discovery phase of any engagement with Dataweavers.


If you’re undecided about which approach to take with the capacity planning or think you might need to change the way you’re operating, reach out to us today and we’ll let you know how we can help.

FREE 90 MINUTE MEETING

Let us show you how fast your Sitecore team can go and get a free licence audit and topology recommendation today.