I can't get Loki to connect to AWS S3 using docker-compose
. Logs are visible in Grafana but the S3 bucket remains empty.
The s3 bucket is public and I have an IAM role attached to allow s3:FullAccess
.
I updated loki to v2.0.0 and changed the period to 24h but it made no difference. There are no errors in the loki logs.
Here are the selected lines from docker logs (loki):
msg="Starting Loki" version="(version=master-4e661cd, branch=master, revision=4e661cde)"
caller=server.go:225 http=[::]:3100 grpc=[::]:9095 msg="server listening on addresses"
caller=worker.go:65 msg="no address specified, not starting worker"
msg="cleaning up mapped rules directory" path=/loki/tmprules
msg=initialising module=memberlist-kv
msg=initialising module=store
msg=initialising module=server
msg=initialising module=ring
msg="value is nil" key=collectors/ring index=1
msg=initialising module=ingester
msg="not loading tokens from file, tokens file path is empty"
msg="instance not found in ring, adding with no tokens" ring=ingester
msg="auto-joining cluster after timeout" ring=ingester
msg=initialising module=table-manager
msg=initialising module=distributor
msg=initialising module=ingester-querier
msg=initialising module=ruler
msg="ruler up and running"
msg="Loki started"
msg="synching tables" expected_tables=132
Here is my loki.config
:
auth_enabled: false
server:
http_listen_port: 3100
distributor:
ring:
kvstore:
store: memberlist
ingester:
lifecycler:
ring:
kvstore:
store: memberlist
replication_factor: 1
final_sleep: 0s
chunk_idle_period: 5m
chunk_retain_period: 30s
schema_config:
configs:
- from: 2020-10-27
store: boltdb-shipper
object_store: s3
schema: v11
index:
prefix: index_
period: 24h
storage_config:
boltdb_shipper:
active_index_directory: /loki/index
cache_location: /loki/index_cache
resync_interval: 5s
shared_store: s3
aws:
s3: s3://AKIARE3@us-east-1/mydomain.com.docker.loki.logs
s3forcepathstyle: true
limits_config:
enforce_metric_name: false
reject_old_samples: true
reject_old_samples_max_age: 168h
Here is docker-compose.yaml
version: "3.8"
networks:
traefik:
external: true
volumes:
data:
services:
fluentd:
image: grafana/fluent-plugin-loki:master
command:
- "fluentd"
- "-v"
- "-p"
- "/fluentd/plugins"
environment:
LOKI_URL: http://loki:3100
LOKI_USERNAME:
LOKI_PASSWORD:
container_name: "fluentd"
restart: always
ports:
- '24224:24224'
networks:
- traefik
volumes:
- type: bind
source: ./config/fluent.conf
target: /fluentd/etc/fluent.conf
logging:
options:
tag: docker.monitoring
loki:
image: grafana/loki:master
container_name: "loki"
restart: always
networks:
- traefik
volumes:
- type: volume
source: data
target: /loki
ports:
- 3100
volumes:
- type: bind
source: ./config/s3.loki.conf
target: /loki/etc/loki.conf
depends_on:
- fluentd
I finally did work this out. It requires a compactor but gives no warning about it. Best practice is to create an AWS s3 bucket without any public access. Next create an IAM user with programmatic access only. Create an access policy which gives full access only to the bucket you created. Attach the policy to the user's permissions. You do not need to attach a policy to the bucket itself. Check if you have "/" in your URL that you escape it with %2F otherwise you will get an auth error. Note that this config is for loki v2.0.0 which was released yesterday.
Here are my complete working docker-compose and loki config files. I put them on an external network to enable prometheus monitoring.
here is my
docker-compose.yaml
Here is my loki config in prometheus/config/s3-loki-bolt-conf.yml. You can name this anything you want but keep the target file name as above as it is the loki default config file.