35Connections through our load balancer reset after exactly 350 seconds of silence. Where do you look and what do you tune?▼mediumJPMorgan ChaseInfosysPhonePe◆ premiumA reset at a suspiciously round number is a fingerprint, not a bug. Matching client keepalives to the lowest timer in the path is the fix; here is the full ladder of knobs.Open full answer →
01A user reports an error. What do 500, 502, 503 and 504 each tell you about where the problem is?▼easyNewCloudflareAccentureInfosysunlockedStatus codes describe an HTTP failure category. Locate the component that emitted the response, then use its logs to distinguish connection, capacity and application failures.Open full answer →
35One app instance keeps flapping out of the load balancer pool, yet between bursts it serves traffic happily. Where do you look?▼mediumLinkedInNetflixAtlassian◆ premiumSelf-healing unavailability is a timeout arithmetic story. Pull the balancer's removal timestamps, lay them against the GC log, and the pause that ejects the instance will be sitting there in plain sight.Open full answer →
63Customers cannot use the app, every health check is green, and the load balancer is happily sending traffic. Find the gap.▼mediumStripeNetflixRazorpay◆ premiumA health endpoint that returns 200 because the process is running tells you the process is running. Interviewers want to hear where you draw the line between a check that pages and a check that removes traffic.Open full answer →
