Monitoring & Troubleshooting Flashcards

What is the purpose of real user monitoring (RUM) in a CDN context?

Answer: Collecting performance data from actual end-user browsers to measure true delivery quality

RUM collects telemetry from real browsers (TTFB, FCP, page load time) to measure actual CDN delivery performance as experienced by users.

A CDN is configured with a 1-hour TTL for static assets, but marketing needs instant cache invalidation capability. Which approach adds that capability without changing TTL?

Answer: Implement cache versioning via URL fingerprinting (e.g., file.abc123.css)

URL fingerprinting embeds a content hash in the filename, making updated files unique URLs that bypass existing cache entries without needing purge operations.

Which CDN metric directly measures the percentage of bytes served from cache versus fetched from origin?

Answer: Offload ratio (byte offload rate)

Byte offload ratio measures what fraction of total data volume was served from CDN cache, reflecting cost savings and origin protection effectiveness.

An HTTP/2 multiplexed connection through a CDN suddenly degrades to HTTP/1.1 behavior. Which log entry would confirm this regression?

Answer: Protocol field showing 'http/1.1' instead of 'h2' in access logs

The protocol field in CDN access logs records whether h2 or http/1.1 was negotiated; a shift to http/1.1 confirms the HTTP/2 multiplexing regression.

A CDN operator receives alerts that origin health checks are failing for one of two origin servers in a load-balanced pool. What is the immediate CDN-level response?

Answer: The CDN automatically removes the unhealthy origin from rotation and routes all traffic to the healthy one

CDN origin health check systems automatically failover by removing unhealthy origins from the load-balancing pool to prevent routing traffic to unavailable servers.

What does a high variance in TTFB measurements across different geographic regions for the same CDN-cached content suggest?

Answer: Inconsistent edge POP performance or uneven POP distribution

For cached content, TTFB should be uniformly low; high regional variance indicates some POPs have performance issues or users are being routed to distant nodes.

Which CDN troubleshooting technique involves sending the same request from multiple geographically distributed vantage points simultaneously?

Answer: Synthetic monitoring with global probes

Synthetic monitoring with globally distributed probes sends identical requests from multiple locations to identify regional performance disparities or routing issues.