Open brsolomon-deloitte opened 1 year ago
I admit that it is a bit ambiguous. The autoscaling policies only applies to nodes that have at least data
or exactly ml
role.
In the example, nodes with roles ["data", "ingest", "transform"]
are data nodes and also ingest and transform nodes, they are not just specialized data
.
nodes.roles | supported? |
---|---|
["data"] |
:heavy_check_mark: |
["data", "ingest" , "transform"] |
:heavy_check_mark: |
["ml"] |
:heavy_check_mark: |
["ingest" , "transform"] |
:x: |
["data", "ml"] |
:x: |
["data", "ml", "ingest" , "transform"] |
:x: |
["ml", "ingest"] |
:x: |
Note that you can't autoscale nodes with roles data
and ml
. There is a validation rule to detect that ML nodes must be in a dedicated autoscaling policy.
I'll see to remove the ambiguity from the doc and maybe add validation for this.
This is still catching folks off-guard. Any movement on amending the docs page on this??
The ECK documentation on
ElasticsearchAutoscaler
is unclear as to which node roles it supports.First it states:
Perhaps I am misreading, but this seems to imply that the autoscaling API, and hence
ElasticsearchAutoscaler
, only supports data nodes or machine learning nodes, and does not support e.g. ingest nodes.But then it gives this example:
...which seems to indicate that nodes other than data nodes are supported.
ECK version:
2.5.0