My Field data cache is full and I want to understand the reason in Marvel. But marvel use field data cache so It has hard time to bring the data. Can I configure Marvel to use docs_value so that it won't use field data cache ?
You can alter the template to use them yourself, I'm not sure it's something we'd directly support at this stage. Though Marvel 2.0 will have this.
How do I alter the template ?
Though I haven't really seen this as a problem before, how many nodes are you monitoring?
5 nodes. But I have heavy use of the field data cache.
We would prefer that third party tools wont interfere with our core business.
I also notices that Marvel use facets. Isn't that feature deprecated?
It is, which is why Marvel 2.0 will move to aggs using KB4.
If it's interfering, then you should really send your Marvel data to a different cluster.
Where can I define it (read from one cluster and write to other cluster)?