I have set up Grafana, Prometheus and loki (2.6.1) as follows on my kubernetes (1.21) cluster:
helm upgrade --install promtail grafana/promtail -n monitoring -f monitoring/promtail.yaml
helm upgrade --install prom prometheus-community/kube-prometheus-stack -n monitoring --values monitoring/prom.yaml
helm upgrade --install loki grafana/loki -n monitoring --values monitoring/loki.yaml
with:
# monitoring/loki.yaml
loki:
schemaConfig:
configs:
- from: 2020-09-07
store: boltdb-shipper
object_store: s3
schema: v11
index:
prefix: loki_index_
period: 24h
storageConfig:
aws:
s3: s3://eu-west-3/cluster-loki-logs
boltdb_shipper:
shared_store: filesystem
active_index_directory: /var/loki/index
cache_location: /var/loki/cache
cache_ttl: 168h
# monitoring/promtail.yaml
config:
serverPort: 80
clients:
- url: http://loki:3100/loki/api/v1/push
# monitoring/prom.yaml
prometheus:
prometheusSpec:
serviceMonitorSelectorNilUsesHelmValues: false
serviceMonitorSelector: {}
serviceMonitorNamespaceSelector:
matchLabels:
monitored: "true"
grafana:
sidecar:
datasources:
defaultDatasourceEnabled: true
additionalDataSources:
- name: Loki
type: loki
url: http://loki.monitoring:3100
I get data from my containers, but, whenever I have a container logging in json format, I can't get access to the nested fields:
{app="product", namespace="api-dev"} | unpack | json
Yields:
My aim is, for example, to filter by log.severity
CodePudding user response:
Actually, following this answer, it occurs to be a promtail
scraping issue.
The current (promtail-6.3.1
/ 2.6.1) helm chart default is to have cri
as pipeline's stage, which expects this kind of logs:
"2019-04-30T02:12:41.8443515Z stdout xx message"
I should have use docker
, which expects json, consequently, my promtail.yaml
changed to:
config:
serverPort: 80
clients:
- url: http://loki:3100/loki/api/v1/push
snippets:
pipelineStages:
- docker: {}