I've been using Hailuo for about six months for customer support triage, and the biggest worry was what happens if their API is down. We can't have tickets just piling up.
So I built a simple failover. If our monitoring pings Hailuo and gets an error, it automatically routes new tickets to a basic rules engine we built in-house. It just categorizes by keywords and assigns to a general queue. It's not smart, but it keeps things moving.
Has anyone else done something like this? I'm curious about how you handle the switch back to Hailuo once it's up again—do you reprocess the tickets or just let the old ones go through the manual flow? 👋
Love this approach! We did something similar at my last place. Your question about handling the switch back is the tricky bit.
We tried reprocessing tickets once, but it created so much duplicate work and confusion for the support team. Our solution was to just let the "rules engine tickets" run their course manually. The failover is meant to be a backup, not a perfect replica.
One extra thing we added - a manual cutoff switch. Sometimes Hailuo would come back online but still be glitchy, and the auto-failback would flip things back and forth. Letting a human decide when to switch back saved us from that chaos. 😅 How are you detecting when Hailuo is truly "up" again? Just the API ping, or something more?