NGINX error metrics should only include 5xx errors

Previously we were also counting 4xx errors, which are not relevant
as a reliability indicator. We may eventually want to restrict this
further to 502/503 (for backend timeouts only).
1 job for master in 5 minutes and 6 seconds (queued for 3 seconds)
Status Job ID Name Coverage
  Test
passed #8880
ai3
run_base_test

00:05:06