This document was translated by GPT-4

1. Introduction

Advanced configuration for DeepFlow Server.

1.1 Custom Deployment Parameters

Although you can define some configurations using the helm --set parameter, we recommend saving your custom configurations in a separate yaml file. For example in values-custom.yaml ：

global:
  storageClass: '<your storageClass>'
  replicas: 1 ## replicas for deepflow-server and clickhouse
  image:
    ## ghcr Image repository address: ghcr.io/deepflowio/deepflow-ce
    ## Dockerhub Image repository address: deepflowce
    ## AliyunYun Image repository address: registry.cn-beijing.aliyuncs.com/deepflow-ce
    repository: registry.cn-beijing.aliyuncs.com/deepflow-ce ## change deepflow image registry to  aliyun

1.2 Use Aliyun Image Repository

global:
  image:
    repository: registry.cn-beijing.aliyuncs.com/deepflow-ce
grafana:
  image:
    repository: registry.cn-beijing.aliyuncs.com/deepflow-ce/grafana

1.3 Modify Server Configuration File

Refer to this server config file (opens new window) to modify the corresponding fields in values, For example, to change the log level:

configmap:
  server.yaml:
    log-level: debug

Subsequent updates can use the -f values-custom.yaml parameter to apply your custom configurations:

helm upgrade deepflow -n deepflow -f values-custom.yaml deepflow/deepflow

1.4 Service Configuration

• global.hostNetwork: Whether to use Host Network, defaulted to false. Use Host Network when CNI network is unavailable, or the traffic collection volume is large.
• global.dnsPolicy: DNS policy for DeepFlow components, defaulted to dnsPolicy. When using Host Network, change to ClusterFirstWithHostNet.
• global.clusterDomain: Cluster domain name, defaulted to cluster.local. If your environment has custom configurations, this needs to be modified. If not modified, it will affect the communication between servers in a multi-server environment.
• global.allInOneLocalStorage: Enable this when it is an all-in-one deployment and local HostPath storage is used, with the default value being false. Data is stored by default in /opt/deepflow-mysql and /opt/deepflow-clickhouse.

1.5 Service Ports

• global.nodePort.deepflowServerIngester
  • The NodePort port number that needs to be exposed by the deepflow-server ingrester module, default is 30033, for agents to transmit data to the server.
  • If there is a conflict, after modifying here, you need to modify analyzer_port in agent-group-config.
• global.nodePort.deepflowServerGrpc
  • The NodePort port number that needs to be exposed by the deepflow-server controller module, default is 30035, for agents to request strategies from the server.
  • If there is a conflict, after modifying here, you need to modify proxy_controller_port in agent-group-config.
• global.nodePort.deepflowServerhealthCheck
  • The HTTP NodePort port number exposed by the deepflow-server controller module to deepflow-ctl, the default is 30417.
  • If there is a conflict, after modifying here you need to specify the --api-port parameter when using the deepflow-ctl command.

1.6 Dependent Services

• global.image.repository: Image repository address for DeepFlow components, default is DockerHub: deepflowce. Domestic users can switch to Aliyun: registry.cn-beijing.aliyuncs.com/deepflow-ce.
• global.image.pullPolicy: Image pull policy for DeepFlow components, default is Always in order to get the ability to update DeepFlow by restarting Pods.
• global.ntpServer: The time synchronization server for DeepFlow, the default value is ntp.aliyun.com.
• global.storageClass: The storageClass used by DeepFlow deployment, the default is empty, which means use the default storageClass.