Best practicess and optimistic values need to keep in order to avoid memory related errors and unexpected cluster outage

Help Workload Profiling
1

Guid me the best practicess and allocation of memeory percentage for below things out of total cluster memory(eg:-If we have leaf nodes. Each leaf node has 64GB ram and 16 cpus. Total Cluster RAM is 128GB and Total Cluster CPU are 32 by excliding master aggregator node.).

What are the best values need to keep inorder to avoid memory related errors and unexpected cluster outager for bleow things.

maximum_memory
maximum_table_memory
Pipeline Resource pool
UI Query read resource Pool
DBA related resource pools

We are creating separate users for PIPELINE,UI & DBA and assign separate resource pool.

2

Please respond to it as soon as possible.

3

You should start with default values. MemSQL comes with a workload manager that will do a good job most of the time. Should you run into resource contention (unlikely) you will need to set up resource pools for each major part of the workload.

4

May I know the default values and share some useful URLS related to workload manager.

5

Here are some helpful links:

Workload Management: https://docs.memsql.com/v7.0/concepts/workload-management/

Default values: https://docs.memsql.com/v7.0/reference/configuration-reference/engine-variables-overview/list-of-engine-variables/#workload-management-sync-variables