I believe we ran into the same issue in our deployment and we ended up redeploying the edge node. Assuming you've setup your edge cluster properly, this should not incur any downtime (we didn't notice anything going down at least).