When I deploy my nextjs app, I get about 30-60 seconds of downtime with blue/green deployments after the new hosts are green. I’m hitting the same URL that is used internally by my healthcheck, and even added a 15s grace period to ensure that the service is properly up. It seems like there is an issue with the Proxy not cutting over in time. Anyone have any clues on this?
I’ve added request logging and it looks like the following happens in order:
(a) New host is brought up. Healthcheck passes.
(b) Old host is destroyed.
(c) Requests are routed to new host.
Effectively, requests aren’t making it to the new host until the proxy flips over. This seems like a serious oversight/bug?
Signed up for premium support and sent an email, we’ll see what the team thinks. If anyone has any tips or suggestions would love to hear them!
Repro repository: GitHub - Optick-Labs/fly-deployment-latency
Thanks for sharing. We’re going to look into this.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.