Add indexer ADR

[jpellizzari]

ADR to explain the complications of labels queries

[enekofb]

⚠️ Disclaimer ⚠️ : I am still in the knowledge-building phase so you might find some nonsense on some of the following :)

I started to look here with the following hypothesis to test out of reading https://blevesearch.com/docs/Index-Mapping/

1) We have more than one document mapping:

• a default document mapping for normalized objects
• X specific document mappings for denormalized objects

2) our normalized object schema is more like a standard kube resources + weave gitops specific fields (like cluster, tenant, etc ..) + unstructured

• searches: via the indexed fields
• visualization: via the unstructured 3) labels belong to the structured schema the same as annotations therefore to the default mapping
• selecting which labels or annotations to index are done via configuration https://github.com/weaveworks/weave-gitops-enterprise/blob/main/pkg/query/configuration/objectkind.go
• the rest are dropped

When do we need mapping extensions?

• One scenario is when we want to index the non generic part of the resource, lets say the .spec.* for example https://fluxcd.io/flux/components/source/ocirepositories/#verification
• In this case we would need to have some specific mapping by kind, but i think that might be the few cases or to implement them as they come?