This is, of course, a band-aid fix for the symptoms, it is not fixing the root cause behind this. I'm aware that there is at least one (probably several) similar places in the tests. Maybe we need some sort of

pihole-FTL check-logs-for "some text here"

which scans FTL.log for exactly this text, probably with some timeout to not wait longer than, say five, seconds. This would then ensure we are not adding sleeps across the tests making them needlessly slow over time.

But I am fully open for better ideas!

This is, of course, a band-aid fix for the symptoms, it is not fixing the root cause behind this.

Is it? The issue is that we execute a process which returns fast but does some stuff in the background and the test run so fast that checking the background stuff a few seconds later will fails because the first process is not done yet.
We need a delay somewhere and I don't think it matters if we add this delay explicitly in the test suite or hide it in some `FTL´ command. Also the amount of sleep time we have in the tests is minimal compared to the overall CI length (2-4 min, > 20 min for RISC). We could probably discuss if I put the sleep at the right place - maybe it would be better placed right before checking the config log rotation.

I did just push something to the branch wait-for-string