We currently do not give a lot of guidance on how beefy clusters should be that are used to deploy our demos.
One notable exception is the Lakehouse demo that has a disclaimer with a few details.
It can be very frustrating for users to spin up a cluster, deploy the demo and notice that nothing works due to memory limitations or similar things.
Can we maybe come up with one shared cluster definition like 5 nodes, 8 cores, 20 gig ram or something like that, which we describe on an extra page along with proposed instance types on the Hyperscalers, and then we just refer to that page from all demos.
We currently do not give a lot of guidance on how beefy clusters should be that are used to deploy our demos. One notable exception is the Lakehouse demo that has a disclaimer with a few details.
It can be very frustrating for users to spin up a cluster, deploy the demo and notice that nothing works due to memory limitations or similar things. Can we maybe come up with one shared cluster definition like 5 nodes, 8 cores, 20 gig ram or something like that, which we describe on an extra page along with proposed instance types on the Hyperscalers, and then we just refer to that page from all demos.