SRE practices: SLOs, incidents, redundancy

Quiz: redundancy

Question 1 of 9

lb-access.log lines with upstream= and ustatus=, a stopped and a sick copy, fail_timeout, proxy_next_upstream with http_500, proxy_read_timeout and proxy_connect_timeout, the sick copy's /healthz, a keepalived floating address moving away and back. The questions follow the lesson “A load balancer and health checks”.

9 questions · threshold 70%

Read the output

What happened to this request?

$ tail -n 1 /var/log/nginx/lb-access.log
::1 [22/Oct/2026:13:11:39 +0000] "GET / HTTP/1.1" 200 upstream=127.0.0.1:8082, 127.0.0.1:8081 ustatus=502, 200 urt=0.000, 0.021 rt=0.021

The client got 502, and nginx logged the 200 for itself.
8082 refused the connection, nginx retried the request on 8081, and the client got 200.
The request went to both copies at once, and nginx picked the good answer.
8082 answered 502, and the client repeated the request itself.

A–D or ↵ Sign in to answer