Hey there, I have a question regarding the topolog...
# announcements
c
Hey there, I have a question regarding the topology of Growthbook (self-hosted solution) - it might be in the documentation, but I might have missed it. ā€¢ Is Growthbook able to run on multiple instances (to ensure high availability), and if so, is that an active/active setup or an active/passive set up?
f
GrowthBook is completely stateless, so you can horizontally scale behind a load balancer without any special setup or configuration.
c
Thank you!
Thank you so much for the reply. I have a few follow up questions: ā€¢ What metrics do we need to observe in order to tell when to scale the Growthbook horizontally or vertically? ā€¢ How about the associated Mongo database? How should we size it in terms of CPU, memory, disk size, disk IO? ā€¢ Are the upgraded docker containers guaranteed to be backward compatible? (That is to say, if we do a rolling upgrade, we should not see breaking changes)
n
Just adding myself to follow up on this thread. šŸ™
šŸ‘ 1
f
Form scaling, memory and CPU are the main metrics. For Mongo, memory is usually the most important factor. We don't store a lot of data, so disk size isn't important. We try to keep upgrades backwards compatible whenever possible. We recommend integrating GrowthBook in a way that removes it from the critical rendering path of your application so if it does need to be taken down for a few minutes occasionally it won't break anything. This involves setting up webhooks and caching feature definitions in something like Redis.
šŸ‘ 1