I am configuring filebeat to send various module output to Elastic Search v7. We are using ILM. I am looking at the ILM settings here: Configure index lifecycle management - ES 7 and I am wondering if someone could clarify a few items. Here is what I do know:
- I know when I run the command
filebeat setup
It loads dashboards, templates, etc. Among these, in the template, it sets an index pattern:
index_patterns" : [
"filebeat-7.6.2-*"
],
and roll-over alias
lifecycle" : {
"name" : "filebeat",
"rollover_alias" : "filebeat-7.6.2"
},
-
I know Kibana's Dashboards are, in some way configured to look at this default index pattern when it searches for fields. Right? I'm not sure how this particular point works.
-
I know when I update File beat it automatically re-loads this information.
-
I know, since I am using ILM, the
index
option in Elasticsearch.output is not available.setup.template.name
andsetup.template.pattern
are also unavailable.
For a subset of machines, I want to place their data in a different index with a different ILM. The mappings,ingest pipelines, and dashboards for the filebeat modules I use will be the same, I just want to re-route them into a separate index so I can apply a different ILM. I'm stuck on how tobest do this. Looking at the ILM documentation, I can see there is an setup.ilm.rollover_alias
option. My questions are:
-
If I define the
setup.ilm.rollover_alias
, will this separate the subset of machines into a different index as well? -
Assuming it does, How does this interact with
filebeat setup
, the Kibana dashboards, and template loading? Will this change be auto-applied every time I update filebeat? -
Assuming it does not, how can I split these up, still use the filebeat & filebeat modules stuff, and ILM?
-
Lastly, if I use the directive
setup.ilm.policy_name
is this telling the template to use a per-existing ilm policy in elastic search, or is this telling filebeat to name the policy it uploads into elastic search this name?