Performance at Scale Flashcards
What is Russian doll caching, and how does it optimize view performance?
Russian doll caching caches nested views, so only the top-level cache invalidates if content changes. This reduces redundant caching and improves speed, especially for complex UIs with nested elements like timelines or feeds.
What is the impact of eager loading in Rails, and when should it be used?
Eager loading loads associated records in a single query, avoiding N+1 query issues. Use includes for associated data in high-traffic areas to reduce database calls and improve load times.
What is a connection pool, and how does it optimize database connections?
Connection pooling limits active database connections, reducing load on the database. Rails manages a pool of connections for reuse, which is critical for apps with high concurrent users to avoid overwhelming the DB.
How can database connection pooling be optimized in a multi-threaded Rails environment?
Set an optimal pool size in database.yml that balances connections with server resources. For multi-threaded environments like Puma, configure a sufficient pool to prevent connection exhaustion while maintaining performance.
Why is connection reaping important for large-scale Rails applications?
Connection reaping closes stale or abandoned connections, preventing pool exhaustion. Configure reaping frequency in database.yml or database adapters to regularly clean up inactive connections, reducing memory usage and avoiding leaks.
How does connection pooling work in multi-threaded vs. multi-process servers?
In multi-threaded servers (e.g., Puma), threads share a connection pool, so set pool size according to the number of threads. For multi-process servers (e.g., Unicorn), each process has its pool, requiring lower pool sizes to avoid exceeding the database connection limit.
How can database connection pooling improve request handling in high-load environments?
Connection pooling reuses existing connections instead of creating new ones for each request, minimizing overhead and latency. This allows faster request processing, essential in high-load environments with billions of users.
How can high traffic spikes be handled in Rails applications?
Use auto-scaling with containers (e.g., Docker on AWS ECS or Kubernetes) or virtual servers. Auto-scaling provisions more servers during peak times and scales down during lulls, preventing overloading and saving costs. Set scaling triggers based on CPU, memory, or request rate metrics.
Why use geo-based load balancing for global traffic distribution?
Geo-based load balancing routes users to the nearest server or region. This reduces latency and balances server load globally, allowing efficient handling of geographically distributed traffic, especially during regional spikes.
What data center considerations are critical for global scaling?
Use a multi-region data center setup to ensure users are served from the closest location, reducing latency and handling regional traffic surges. AWS and Google Cloud offer managed multi-region setups that simplify global scaling.
What is request batching, and how does it help during traffic spikes?
Batching groups multiple user requests (e.g., contest entries) and processes them as a single request, reducing the number of direct database calls. This strategy minimizes database load during spikes and helps manage high volumes efficiently.
