We saw test failure in a CI run, because the time between the first `ping` and the last `pong` was about 4 seconds and the response for `getnettotals` was not up to date after one second
.
Increasing the first timeout from 1 s to 10 s should allow the nodes to synchronize their states and make the test more reliable when it is run on a slow system.