Performance, Caching, and Dispatcher Configuration
Performance questions combine caching strategy with Dispatcher rules. A candidate must understand the Dispatcher as a caching reverse proxy: cache rules, filter rules, and farm configuration. Know the difference between author and publish dispatcher setups, the role of the cache invalidation agent (flush), and how stateless versus stateful headers affect caching. Understand TTL-based caching, the /statistics endpoint, and how to invalidate by path or by activation.
Key concepts include client-side versus server-side caching, the role of the CDN in front of the dispatcher on Cloud Service, and how cache headers (Cache-Control, Surrogate-Control) influence behavior. Know the performance implications of component design: number of requests, client library bundling, lazy loading, and the cost of deep content trees. Understand the query performance tools (explain query, index definitions) and why unbounded queries kill performance.
Traps include assuming everything is cached by default (it is not), or forgetting that query strings and cookies can bypass the cache. Another trap is a filter rule that is too permissive, exposing admin paths, or too restrictive, blocking legitimate API calls. Expect questions on why a page is not cached (dispatcher rules, headers, or authentication) and on the difference between invalidating the dispatcher and invalidating the CDN. Also watch for confusion between author-tier and publish-tier caching.
Rule of thumb: cache aggressively at the edge, invalidate precisely by path, and keep the filter allowlist minimal. Worked example: a page with a personalized header is not caching. Move personalization client-side or to a separate uncached fragment so the main page remains cacheable.