According to the production checklist, a minimum of 1000 IOPS is required for Dgraph, while 3000 is recommend for medium/large data sets. I can see this being the case for Alpha nodes, but would this apply to Zero instances as well, and why?
https://dgraph.io/docs/deploy/production-checklist/
Thanks in advance!
Shouldn’t be needed. Zero’s need a fraction of the resources of alphas, which do the heavy lifting.
Thanks - that’s what I thought as well. From your experience, what’s a reasonable IOPS rate for Zero instances supporting a high-throughput system?
Thanks again!