Hello, I’m curious about how you typically set up ...
# community-help
a
Hello, I’m curious about how you typically set up environments in Typesense. Our usual approach is to have separate environments for development (dev), testing (tst), and production (prd). I’d like to understand what you recommend in this case: • Would you suggest using a single cluster with multiple collections for each environment (dev, tst, prd)? My concern here is that if the dev collection is heavily tested, misused, or even broken, it might impact the production environment. • Alternatively, would it be better to keep production in a separate cluster, and group dev and tst together? • Or is using three completely separate clusters (dev, tst, prd) considered best practice, despite the higher cost? Additionally, I understand that storing similar or identical data within the same cluster may be more space-efficient. Could you confirm if that’s correct? Thanks in advance for your insights!
f
Hey @Antoon Van Dessel , Using multiple clusters for different environments does lead to a higher total cost, but the impact will be less noticeable as your dataset grows. For example, if a cluster with 0.5GB RAM can comfortably fit your dataset multiple times over, you would essentially be leaving RAM on the table. Same goes for high traffic clusters with smaller datasets, where you have to upgrade RAM in order to get access to larger-sized CPU configurations, and the headroom of RAM isn't used totally. For larger-scaled clusters, this won't be as noticeable though. As for your environment setup, having completely isolated clusters for different environments is by far the safest choice, and one that most of our users actually take advantage of. It prevents accidental issues on production, like indexing changes, preset changes and some curation rules being updated when they shouldn't be. Non-production clusters can have High Availability disabled, which is going to reduce runtime costs by more than 200%, as a single-node cluster is less than a third of the price of a Highly Available one. Storing similar or identical data within the same cluster can be more space-efficient due to how Typesense handles data, but careful consideration is needed depending on your use case and risk-averseness for impacting production.
a
Okey thank you that is a clear answer!
🙌 1