Loki Helm configuration ( AWS S3 with IRSA) https://medium.com/techlogs/grafana-loki-with-aws-s3-backend-through-irsa-in-aws-kubernetes-cluster-93577dc482a

loki_grafana_values.yaml

loki:
  auth_enabled: false
  commonConfig:
    path_prefix: /var/loki
    replication_factor: 1
  compactor:
    apply_retention_interval: 1h
    compaction_interval: 5m
    retention_delete_worker_count: 500
    retention_enabled: true
    shared_store: s3
    working_directory: /data/compactor
  config:  
      schema_config:
        configs:
        - from: 2020-05-15
          store: boltdb-shipper
          object_store: s3
          schema: v11
          index:
            period: 24h
            prefix: loki_index_
        
      storage_config:
        aws:
          region: eu-central-1
          bucketnames: alex-loki-test
          s3forcepathstyle: false
          #s3forcepathstyle: true  <-- This is the main culprit; comment it out ? -? https://github.com/grafana/loki/issues/7024
        boltdb_shipper:
          shared_store: s3
          cache_ttl: 24h
      limits_config:
          max_entries_limit_per_query: 50000 # default is 5000
  serviceAccount:
    create: true
    name: loki-sa
    annotations:
       eks.amazonaws.com/role-arn: "arn:aws:iam::xxx:role/loki_s3_role"
  write:
     replicas: 2
  read:
    replicas: 1

grafana:
    enabled: true
    sidecar:
      datasources:
        enabled: true
    image:
      tag:  7.5.17
    grafana.ini:
      users:
        default_theme: dark

from where fake folder name is getting to s3?

@shreysway fake is the default name for a single-tenant id. This folder will be created in s3 by default if you aren't running in muti-tenat mode and your chunks will live there https://grafana.com/docs/loki/latest/operations/multi-tenancy/