Loki

Loki #

部署 #

部署模式 #

loki的helm chart的values.yaml中的deploymentMode配置部署模式。

有 3 种选项：

  • SingleBinary：Loki 作为单一二进制文件部署，适用于小型安装，通常没有高可用性，支持每天几十 GB 的数据。
  • SimpleScalable(默认值）：Loki 作为3个目标部署：读取(loki-read)、写入(loki-write)和后端(loki-backend)。适用于中型安装，比分布式部署更易于管理，支持每天约1TB的数据。
  • Distributed：Loki作为独立的微服务部署。最复杂但功能最强，适用于大型安装，通常超过每天1TB的数据。

还有 2 种附加模式，用于在部署模式之间迁移：

  • SingleBinary<->SimpleScalable：从SingleBinary迁移到 SimpleScalable（或反向迁移）
  • SimpleScalable<->Distributed：从SimpleScalable迁移到 Distributed（或反向迁移）

注意：SimpleScalable 和 Distributed 要求使用对象存储。

SimpleScalable部署 #

定制的values.yaml:

  1imagePullSecrets:
  2  - name: regsecret
  3
  4minio:
  5  enabled: false
  6
  7lokiCanary:
  8  enabled: false
  9resultsCache:
 10  enabled: false
 11chunksCache:
 12  enabled: false
 13
 14loki:
 15  image:
 16    registry: registry.frognew.com
 17    repository: library/grafana/loki
 18  storage:
 19    bucketNames:
 20      chunks: loki-chunks
 21      ruler: loki-ruler
 22      admin: loki-admin
 23    type: s3
 24    s3:
 25      endpoint: http://rook-ceph-rgw-ceph-object-store.rook-ceph.svc.cluster.local:80
 26      region: unknown
 27      accessKeyId: <accessKeyId>
 28      secretAccessKey: <secretAccessKey>
 29      s3ForcePathStyle: true
 30      insecure: false
 31  compactor:
 32    delete_request_store: s3
 33    compaction_interval: 10m
 34    retention_enabled: true
 35    retention_delete_delay: 2h
 36    retention_delete_worker_count: 150
 37  limits_config:
 38    max_cache_freshness_per_query: 10m
 39    reject_old_samples: true
 40    reject_old_samples_max_age: 1w
 41    split_queries_by_interval: 15m
 42    retention_period: 365d
 43  tenants:
 44    - name: <tenants-name>
 45      password: <tenants-password>
 46  analytics:
 47    reporting_enabled: false
 48
 49  schemaConfig:
 50    configs:
 51      # https://grafana.com/docs/loki/latest/operations/storage/tsdb/
 52      - from: 2024-05-14
 53        store: tsdb
 54        object_store: s3
 55        schema: v13
 56        index:
 57          prefix: loki_index_
 58          period: 24h
 59  storage_config:
 60    hedging:
 61      at: 250ms
 62      max_per_second: 20
 63      up_to: 3
 64
 65gateway:
 66  enabled: false
 67
 68read:
 69  replicas: 3
 70write:
 71  replicas: 3
 72  persistence:
 73    size: 50Gi
 74    storageClass: local-storage
 75backend:
 76  replicas: 3
 77  persistence:
 78    size: 50Gi
 79    storageClass: local-storage
 80
 81test:
 82  enabled: false
 83
 84ingress:
 85  enabled: false
 86
 87monitoring:
 88  dashboards:
 89    enabled: false
 90  rules:
 91    enabled: false
 92  serviceMonitor:
 93    enabled: false
 94  selfMonitoring:
 95    enabled: false
 96    grafanaAgent:
 97      installOperator: false
 98  lokiCanary:
 99    enabled: false
100
101sidecar:
102  rules:
103    enabled: false

更多关于SimpleScalable部署的内容，可参考“Grafana Loki笔记02: 使用Helm安装Loki”

SingleBinary部署 #

values.yaml:

  1imagePullSecrets:
  2  - name: regsecret
  3
  4minio:
  5  enabled: false
  6
  7lokiCanary:
  8  enabled: false
  9resultsCache:
 10  enabled: false
 11chunksCache:
 12  enabled: false
 13
 14loki:
 15  image:
 16    registry: registry.frognew.com
 17    repository: library/grafana/loki
 18  storage:
 19    bucketNames:
 20      chunks: loki-chunks
 21      ruler: loki-ruler
 22      admin: loki-admin
 23    type: s3
 24    s3:
 25      endpoint: http://rook-ceph-rgw-ceph-object-store.rook-ceph.svc.cluster.local:80
 26      region: unknown
 27      accessKeyId: <accessKeyId>
 28      secretAccessKey: <secretAccessKey>
 29      s3ForcePathStyle: true
 30      insecure: false
 31  compactor:
 32    delete_request_store: s3
 33    compaction_interval: 10m
 34    retention_enabled: true
 35    retention_delete_delay: 2h
 36    retention_delete_worker_count: 150
 37  limits_config:
 38    max_cache_freshness_per_query: 10m
 39    reject_old_samples: true
 40    reject_old_samples_max_age: 1w
 41    split_queries_by_interval: 15m
 42    retention_period: 365d
 43  tenants:
 44    - name: <tenant-name>
 45      password: <tenant-password>
 46  analytics:
 47    reporting_enabled: false
 48  commonConfig:
 49    replication_factor: 1
 50  schemaConfig:
 51    configs:
 52      # https://grafana.com/docs/loki/latest/operations/storage/tsdb/
 53      - from: 2024-05-14
 54        store: tsdb
 55        object_store: s3
 56        schema: v13
 57        index:
 58          prefix: loki_index_
 59          period: 24h
 60  ingester:
 61    chunk_encoding: snappy
 62  tracing:
 63    enabled: true
 64  querier:
 65    # Default is 4, if you have enough memory and CPU you can increase, reduce if OOMing
 66    max_concurrent: 2
 67  storage_config:
 68    hedging:
 69      at: 250ms
 70      max_per_second: 20
 71      up_to: 3
 72
 73gateway:
 74  enabled: false
 75
 76deploymentMode: SingleBinary
 77singleBinary:
 78  replicas: 1
 79  resources:
 80    limits:
 81      cpu: 3
 82      memory: 2Gi
 83    requests:
 84      cpu: 500m
 85      memory: 500Mi
 86  extraEnv:
 87    # Keep a little bit lower than memory limits
 88    - name: GOMEMLIMIT
 89      value: 1875MiB
 90
 91# Zero out replica counts of other deployment modes
 92backend:
 93  replicas: 0
 94read:
 95  replicas: 0
 96write:
 97  replicas: 0
 98
 99ingester:
100  replicas: 0
101querier:
102  replicas: 0
103queryFrontend:
104  replicas: 0
105queryScheduler:
106  replicas: 0
107distributor:
108  replicas: 0
109compactor:
110  replicas: 0
111indexGateway:
112  replicas: 0
113bloomCompactor:
114  replicas: 0
115bloomGateway:
116  replicas: 0
117
118test:
119  enabled: false
120
121ingress:
122  enabled: false
123
124monitoring:
125  dashboards:
126    enabled: false
127  rules:
128    enabled: false
129  serviceMonitor:
130    enabled: false
131  selfMonitoring:
132    enabled: false
133    grafanaAgent:
134      installOperator: false
135  lokiCanary:
136    enabled: false
137
138sidecar:
139  rules:
140    enabled: false

Archives #

© 2024 青蛙小白 | 总访问量 | 总访客数