Module 14 · Network debugging · Drill

Network debugging drill

Scenario prompts that check whether network debugging is usable under pressure. Answer out loud, then compare.

4 prompts25 minNew

0 of 4 graded

Prompt 1
Checkout's client-side metrics show p99 latency jumping from 300 ms to 2.5 s, while the payment service's own logs show every request handled in under 80 ms. Both teams say the problem is the other's. Lay out how you would settle it with measurements and captures, and what each possible finding implies.
Say your answer out loud first. Thirty seconds is enough.
Prompt 2
Every Monday morning the first request from each API instance to the database fails with 'connection reset by peer', and retries succeed. Separately, a newly added network ACL makes some clients hang right after connecting. Explain both with firewall and state-table behaviour and fix them.
Say your answer out loud first. Thirty seconds is enough.
Prompt 3
Your video-calling feature connects directly peer to peer for most users, but 15 to 20 percent of calls fail to connect, concentrated among corporate and some mobile users. Explain why with NAT and firewall behaviour, and design the connectivity path so nearly all calls succeed.
Say your answer out loud first. Thirty seconds is enough.
Prompt 4
During a regional incident, a dependency slowed from 50 ms to 900 ms at p99 for ten minutes, and the calling service, which has three layers of callers above it, went fully down for an hour. Reconstruct how a short network or dependency slowdown became a long outage, and specify the timeout, retry and limit settings you would put in place.
Say your answer out loud first. Thirty seconds is enough.