Zero Downtime Deploy
Is there a way to do a zero downtime deploy such that new instances are started and waiting and traffic sent over after a healthcheck of somekind? The deploys are pretty good as is but we do get a few seconds of failed connections after a deploy. We are hosting an API so missing those API calls during that time is not great. Not sure if it's something we are not doing propperly or if there isn't a way to facilitate that at this time.
8 Replies
Project ID:
N/ASolution
ah ... thank you for the prompt response ... I scrolled right past that π€¦ββοΈ
no prob π
do you have a volume on your service though?
meaning do we have a lot of req/sec?
oh ... no
I was reading that as "high volume" on the service
ah gotcha, i only asked because having a volume will guarantee downtime regardless of if you have a healthcheck