v1.4.0

Important

• Backup jobs against MariaDB 11+ work again. SrvKit now detects mariadb-dump or mysqldump inside the container on every run, so a container image upgrade or DB migration needs no reconfiguration. Jobs on MariaDB ≥ 11 had been failing with a bare exited 127 since their migration; when neither binary is present the error now names both candidates.
• A run that produced no content is a failure, not a green row. Every job type measures its raw content (dump stream, copied database, packed sources) and a 0-byte run is marked FAILED — aborting before the upload, so an empty dump can never overwrite the last good backup at the target.
• A job that stays failed is re-announced. New global setting in Settings → General (default every 24 h, Off restores the previous one-alert-then-silence behaviour). The reminder says how long the job has been failing, how many runs that covers, and when a good backup last landed. The interval is measured from the last alert, so a job failing every 15 minutes still sends at most one message per interval.

Other

• Job rows show the size of the last run's content, so a dump that is shrinking rather than empty is visible too.
• The MySQL job info box now states that both MariaDB and MySQL are supported.

Upgrade note

Re-run each MySQL-type job once manually (▶) after deploying and confirm a non-empty dump lands in the target. Treat the first successful run as the recovery point, not the job status going green.

A Files job whose sources are all zero-byte will now report FAILED — intended, but a behaviour change on existing jobs.