High Availability (HA) for Redis
D
David Hummel
I would be happy to pay for two (or more) matching sized Redis instances if they were offered as a option for anyone interested in outage resiliency. You could guarantee them to be on different hosts (or optionally regions), automatically configure them as a redundant Redis cluster, and everyone wins!
Log In
R
Rich Shakespeare
Thanks for tracking this. We've recently experienced downtime due to no HA on Redis, during maintenance. Given the plans can get very expensive and apps depend on Redis 24/7, it seems highly necessary that during maintenance there is failover, otherwise this offering isn't commercially usable for apps which require no downtime.
We unfortunately will have to not use it until this is implemented at least for maintenance.
a
ankit
We use Redis for queues and at high throughput the 1 second lag to write-to-disk could be problematic. It is not clear if the current write-to-disk is executed every second, guaranteed completes within the second, or both.
Anurag Goel
marked this post as
open
Anurag Goel
marked this post as
planned