How one hanging API takes down a whole Spring Boot app
We pointed a Spring Boot app at a partner API that never answers. Defaults: offline in 10 seconds, no recovery. Timeouts fixed it; virtual threads and Apache HttpClient each had a surprise.
Engineering notes.
Tested write-ups on the problems that surface once software meets production: what breaks, why, and the fix, with the code and the numbers.
Why your Spring Boot @Scheduled job runs twice (and what actually stops it)
Add a second instance and every @Scheduled job runs twice. We counted the duplicates across six setups: ShedLock fixes it, but only with two settings that are easy to get wrong.
The case of the vanishing JobRunr runs
Spring, Quartz and JobRunr ran the same job every minute, and JobRunr quietly lost a run. The log line that gave it away, the gap behind it and the 5.2.0 fix.
Going, going, gone: closing Firebase auctions on time
Bids pile up in the last seconds, and a once-a-minute sweep closes auctions late. The pattern we recommend: the deadline in the bid transaction, one Cloud Task per auction and an idempotent close.