A simple bash script to dump all postgresql databases in a running podman container and send ntfy notifications. This should also work with docker, just replace podman with docker in the script (untested). pg_dumpall_ntfy.sh
The script does not do comprehensive checking, but can be run manually or automated (eg CRON). I use this script to do database dumps on an independent schedule, and to place the backups in a directory that is subsequently backed up (I use docker-borgmatic in a podman pod). Setting a suitable retention (automation interval, plus BACKUP_ROTATE setting) allows for multiple database dumps to be retained per backup run. Requires adding appropriate specific settings to the environment variables listed at the head of the script.
The dump command podman exec -i $PG_CONTAINER /bin/bash -c "PGPASSWORD=$DB_PASSWORD pg_dumpall --username $DB_USER" | gzip > "$backup_file" includes writing to a file, it will complete successfully as long as the file is written - Not necessarily because the dump has worked. The notification on complete therefore notifies of the current database dump filesize. I monitor these through the ntfy notifications, and periodically check the dump files.
This script was created as a quick and dirty workaround because docker-borgmatic does not currently support postgresql 16 as per issue pg_dump does not support postgresql 16..