Loki in Production📜
This chart supports running Loki in two architecture modes, monolith and simple-scalable. It is strongly encouraged to utilize the simple-scalable architecture in production environments as it allows more fine grained control over scaling of read/write nodes compared to the monolith.
Big Bang Simple Scalable Recommended Production Configuration📜
This example assumes external dependencies of: * AWS S3
loki:
strategy: "scalable"
objectStorage:
#If using Grafana enterprise, dropping the `https://` is required
endpoint: https://s3.us-gov-west-1.amazonaws.com
region: us-gov-west-1
bucketnames:
# Loki since their 1.0 Chart release allows you to specify separate buckets.
chunks: loki-logs #For storage of log data in 'chunks'
ruler: loki-logs #For Loki configured rules https://grafana.com/docs/loki/latest/rules/
admin: loki-admin #For Grafana Enterprise configuration/persistence only
deletion: loki-deletion #The storage bucket that stores the delete requests
values:
loki:
memcached:
# configure how Loki will cache requests, chunks, and the index to a backing cach store.
chunk_cache:
enabled: true
host: "elasticache.XXXX:6379"
service: "memcached"
batch_size: 256
parallelism: 10
write:
replicas: 3
persistence:
size: 50Gi
read:
replicas: 5
persistence:
size: 50Gi
Cloud Credential Configuration📜
Loki is able to use IAM Profiles attached to nodes on your instances, if you don’t specify loki.objectStorage.accessKey
or loki.objectStorage.accessSecret
loki will use the Profile attached to the instance. If you specify accessKey
& accessSecret
they will be mounted within a non encrypted/obfuscated configMap via the loki-simple-scalable chart so it is encouraged to utilize instance profiles where possible.
AWS IRSA Configuration📜
To use AWS IRSA within Loki, ensure the following values are set along with the usual loki.objectStorage
values via BigBang:
loki:
values:
loki:
storage:
s3:
endpoint: ""
serviceAccount:
annotations:
eks.amazonaws.com/role-arn: "my-role-arn"
Override A Custom Configuration📜
If the above recommended configuration is not enough, you would like to add in additional options or utilize something like AWS DynamoDB as a table-manager instead of tsdb-shipper, you can override in your own full configuration under the global.config
value:
loki:
values:
loki:
config: |
auth_enabled: true
server:
http_listen_port: 3100
log_level: debug
http_server_read_timeout: 300s
http_server_write_timeout: 300s
...
storage_config:
aws:
s3: s3://access_key:secret_access_key@region/bucket_name
dynamodb:
dynamodb_url: dynamodb://access_key:secret_access_key@region
schema_config:
configs:
- from: 2022-01-01 # Anything in the past
store: aws
object_store: s3
schema: v12
index:
prefix: prefix_loki_logs_
period: 24h
Monolith Recommended Configuration📜
If you wish to use the monolith chart in production it is recommended that you instead utilize the simple-scalable-chart and external object storage, but you can deploy the monolith installation which is a single loki pod which writes/stores to local PVC storage:
loki:
strategy: monolith
values:
singleBinary:
persistence:
size: 40Gi
New Config Value Options📜
Since Big Bang chart version 4.X, Loki-simple-scalable version 1.X there is the ability to configure some options via loki
value definitions, these options include:
| Value | Documentation |
| ------|---------------|
| storage_config: {}
| https://grafana.com/docs/loki/latest/configuration/#storage_config |
| schemaConfig: {}
| https://grafana.com/docs/loki/latest/configuration/#schema_config |
| structuredConfig: {}
| Structured loki configuration, takes precedence over global.config
, global.schemaConfig
, global.storageConfig
|
| query_scheduler: {}
| https://grafana.com/docs/loki/latest/configuration/#query_scheduler |
| commonConfig: {}
| https://grafana.com/docs/loki/latest/configuration/#common |
| memcached.chunk_cache: {}
| https://grafana.com/docs/loki/latest/configuration/#cache_config |
| memcached.results_cache: {}
| https://grafana.com/docs/loki/latest/configuration/#query_range |
An example of using these new values is the memcached
example in the Simple Scalable Recommended Production Configuration section above.