Content revised 21 October 2026.
A traffic spike can expose a limit in hosting, application work or third-party services. Prepare the important customer journey before a major campaign rather than waiting for the homepage to fail.
Identify the critical path
List the pages and actions that must remain available: landing pages, forms, account functions or checkout. A cached article and a transaction requiring database work can place very different demands on the system. Expected visit totals alone don't describe the workload.
Agree capacity with the host
Review the resources, limits and scaling process that apply to the actual plan. Ask what happens when a limit is reached and who receives alerts. Extra storage or a larger monthly transfer allowance doesn't necessarily fix a slow database or overloaded application worker.
Prepare a controlled test
- Get authorisation for the test environment and traffic level.
- Use representative journeys and realistic assumptions.
- Monitor errors, response times and resource use.
- Define stop conditions and a recovery plan.
Keep caching appropriate
Caching public resources can reduce repeated work, but personal or transaction data needs careful treatment. Check forms and updated content after configuration changes. Do not cache everything indiscriminately just to improve a synthetic result.
Plan the response
Assign who watches the launch, who can change capacity and how customer-facing issues are handled. Review the findings afterward and prioritise demonstrated bottlenecks. Preparation reduces uncertainty, but it is not a promise that any volume of traffic can be handled without limits.
Official guidance: Google’s context for server capacity and crawling.

