Splunk doesn't have a high-availability solution for apps. You can stand up a warm standby Splunk instance with DB Connect installed on it and use rsync (or a similar utility) to copy data from the primary to the standby. I don't know the exact files you'd have to synchronize. The takeover process would be manual or automated outside of Splunk. Likewise, when the original server comes back on-line you will have to take steps to ensure only one instance performs inputs or you'll risk duplicate data.
--- If this reply helps you, an upvote would be appreciated.