

2·
14 days agosystemd timers are definitely nicer than cron for this (and basically all cron jobs). there’s both a restart on failure with backoff if you want it to handle the retry, or a way to schedule it based on the last exit time so the script can manage the backoff like you are doing now
on the last part, page numbers are hard to trust. I assume you don’t have a choice, but that’s why I prefer next page tokens when designing an api instead of numbered pages. then you know the data you’re getting is exactly after the previous data
this entire account seems to be automated llm slop posts, not sure what the goal is