How to force long term retention of logs in loki with tsdb

Grafana Loki
1

I’ve been tinkering around a couple of days with my Loki setup handling logs through promtail and using s3 for long term backups of the kubernetes logs.

I’m a bit confuse since the two places I’ve found referencing to that says that Loki will by default not have a retention period set, implying that it will store them forever.
Is my interpretation correct? Is this even true with recent v11 tsdb and S3 as backend?
I’m installing loki-stack through helm chart, which means the loki final config after templating is as folows

auth_enabled: false
chunk_store_config:
  max_look_back_period: 0s
compactor:
  retention_enabled: true
  shared_store: s3
  working_directory: /data/loki/boltdb-shipper-compactor
ingester:
  chunk_block_size: 262144
  chunk_idle_period: 3m
  chunk_retain_period: 1m
  lifecycler:
    ring:
      replication_factor: 1
  max_transfer_retries: 0
  wal:
    dir: /data/loki/wal
limits_config:
  enforce_metric_name: false
  ingestion_burst_size_mb: 24
  ingestion_rate_mb: 50
  max_entries_limit_per_query: 5000
  max_global_streams_per_user: 10000
  per_stream_rate_limit: 10MB
  reject_old_samples: true
  reject_old_samples_max_age: 168h
memberlist:
  join_members:
  - 'loki-memberlist'
schema_config:
  configs:
  - from: "2023-01-05"
    index:
      period: 24h
      prefix: index_
    object_store: s3
    schema: v11
    store: tsdb
server:
  grpc_listen_port: 9095
  grpc_server_max_recv_msg_size: 54857600
  grpc_server_max_send_msg_size: 54857600
  http_listen_port: 3100
storage_config:
  aws:
    bucketnames: k8s-longterm-logs
    s3forcepathstyle: false
  boltdb_shipper:
    active_index_directory: /data/loki/boltdb-shipper-active
    cache_location: /data/loki/boltdb-shipper-cache
    cache_ttl: 24h
    shared_store: filesystem
  filesystem:
    directory: /data/loki/chunks
  tsdb_shipper:
    active_index_directory: /data/loki/index
    cache_location: /data/loki/index_cache
    cache_ttl: 24h
table_manager:
  retention_deletes_enabled: false
  retention_period: 0s

As you can see, the helm chart templates on loki-stack still populates some boltdb entries, which I’m not 100% are used at all, but I’m more worried about the remaining config parameters being compatible with my needs.

I need to use s3 as data backend since almost everything else gets destroyed and recreated eventually, meaning s3 is the only thing I can trust for long term availability.

On the same direction, do you know if once loki objects are uploaded into S3, those files are susceptible to modification or deletion, assuming a “infinite time retention” ?

2

Please see Log retention | Grafana Loki documentation. regarding Loki retention.