Collect tke logs through daemonset CRD

stay TKE in LogListener In order to DaemonSet mode , It can be done by CRD Method to create a collection configuration collection TKE Cluster log . How does this article describe CRD Method to create a collection configuration .

Prerequisite

You need to enable the log collection function , Please refer to Start log collection

Create a collection configuration

You just need to define LogConfig CRD You can create an acquisition configuration ,LogListener according to LogConfig CRD Changes to modify the corresponding logging service CLS Journal topics , And set the bound machine group .CRD The format is as follows ：

apiVersion: cls.cloud.tencent.com/v1
   kind: LogConfig                          ##  The default value is 
metadata:
  name: test                                ## CRD Resource name , The only one in the cluster 
spec:
  clsDetail:
    topicId: xxxxxx-xx-xx-xx-xxxxxxxx       ## CLS Log subject ID, Log topics need to be in CLS Create in advance , And is not occupied by other acquisition configurations 
    logType: minimalist_log                 ##  Log collection format ,json_log representative  json  Format ,delimiter_log Represents the separator format ,minimalist_log Represents single line full-text format ,multiline_log Represents multi line full-text format ,fullregex_log Stands for fully regular format 
    extractRule:                            ##  extract 、 Filtering rules 
      ...
  inputDetail:
    type: container_stdout                  ##  Type of collection log , Include container_stdout（ Container standard output ）、container_file（ Container file ）、host_file（ Host files ）

    containerStdout:                        ##  Container standard output 
      namespace: default                    ##  Of the collection container kubernetes Namespace , If you don't specify , Represents all namespaces 
      allContainers: false                  ##  Whether to collect the standard output of all containers in the specified namespace 
      container: xxx                        ##  Satisfy includeLabels Of Pod Container name in , Only in the designated includeLabels When using 
     includeLabels:                         ##  The acquisition contains the specified label Of Pod
        k8s-app: xxx                        ##  Collect only pod Configuration in label "k8s-app=xxx" Of pod Generated log , And workloads、allContainers=true Cannot specify at the same time 
      workloads:                            ##  Of the container to be collected Pod Of kubernetes workload
      - namespace: prod                     ## workload The namespace of 
        name: sample-app                    ## workload Name 
        kind: deployment                    ## workload type , Support deployment、daemonset、statefulset、job、cronjob
        container: xxx                      ##  Name of the container to collect , If you don't specify , representative workload Pod All containers in 

    containerFile:                          ##  Documents in the container 
      namespace: default                    ##  Of the collection container kubernetes Namespace 
      container: xxx                        ##  Collection container name 
     includeLabels:                         ##  The acquisition contains the specified label Of Pod
        k8s-app: xxx                        ##  Collect only pod Configuration in label "k8s-app=xxx" Of pod Generated log , And workload Cannot specify at the same time 
      workload:                             ##  Of the container to be collected Pod Of kubernetes workload
        name: sample-app                    ## workload Name                   
        kind: deployment                    ## workload type , Support deployment、daemonset、statefulset、job、cronjob
      logPath: /opt/logs                    ##  Log folder , Wildcard characters are not supported 
      filePattern: app_*.log                ##  Log file name , Support for wildcards  *  and  ? ,*  Indicates matching multiple arbitrary characters ,?  Means to match a single arbitrary character 

    hostFile:                               ##  Host files 
      logPath: /opt/logs                    ##  Log folder , Support for wildcards 
      filePattern: app_*.log                ##  Log file name , Support for wildcards  *  and  ? ,*  Indicates matching multiple arbitrary characters ,?  Means to match a single arbitrary character 
      customLablels
        k1: v1

View collection configuration

• View all collection configurations

[[email protected] ~]# kubectl get logconfigs
NAME            AGE
cls-delimiter   48m
stdout-lgs      4h21m
www-nginx       75m

• View the details and status of the collection configuration

1. among stdout-lgs Is the name of the configuration , It can be replaced according to the actual situation ;
2. status The field values for Synced Indicates that the application configuration is successful , Otherwise, the application configuration fails .

[[email protected] ~]# kubectl get logconfigs stdout-lgs -o yaml
apiVersion: cls.cloud.tencent.com/v1
kind: LogConfig
metadata:
  creationTimestamp: "2021-04-07T09:33:42Z"
  generation: 2
  managedFields:
  - apiVersion: cls.cloud.tencent.com/v1
  ......
  name: stdout-lgs
spec:
  clsDetail:
    extractRule:
      beginningRegex: (\S+)\s-[^-]+(\S+)\s\[([^\]]+)\]\s"(\w+)\s(\S+)\s([^"]+)"\s(\d+)\s(\d+)\s(\S+)[^-]+([^"]+)"\s"([^"]+)[^-]+([^"]+)"\s"([^"]+).*
      keys:
      - client_ip
      - remote_user
      - request_time
      - request_method
      - request_url
      - http_protocol
      - status_code
      - body_bytes_sent
      - handle_time
      - http_referer
      - http_user_agent
      - real_ip
      - log_time
      logRegex: (\S+)\s-[^-]+(\S+)\s\[([^\]]+)\]\s"(\w+)\s(\S+)\s([^"]+)"\s(\d+)\s(\d+)\s(\S+)[^-]+([^"]+)"\s"([^"]+)[^-]+([^"]+)"\s"([^"]+).*
    logType: fullregex_log
    topicId: 4511efaa-2cad-4c3c-8c7c-d3bb5a58d903
  inputDetail:
    containerStdout:
      includeLabels:
        k8s-app: tke-cls-nginx-std
      namespace: cls-wk
    type: container_stdout
status:
  status: Synced
  

Sample collection

standard output

• Collect one namespace yes cls-wk, The label is tke-cls-nginx Configuration of standard output of all containers

apiVersion: cls.cloud.tencent.com/v1
kind: LogConfig
metadata:
    #  Configuration name , Cluster unique 
    name: test123
spec:
    #  Consumer configuration 
   clsDetail:
       #  The consumption of topic id
       topicId: 4224b4f3-9ee8-46e3-822a-xxxxxxxxx
       #  Extraction mode 
       logType: minimalist_log
   inputDetail:
       #  Collection type 
       type: container_stdout
       containerStdout:
           #  Belongs to namespace
           namespace: cls-wk
           #  Label corresponding to the container 
           includeLabels:
               k8s-app: tke-cls-nginx

• Regular expression extraction pattern

apiVersion: cls.cloud.tencent.com/v1
kind: LogConfig
metadata:
    #  Configuration name , Cluster unique 
    name: stdout-lgs
spec:
    clsDetail:
        extractRule:
            beginningRegex: (\S+)\s-[^-]+(\S+)\s\[([^\]]+)\]\s"(\w+)\s(\S+)\s([^"]+)"\s(\d+)\s(\d+)\s(\S+)[^-]+([^"]+)"\s"([^"]+)[^-]+([^"]+)"\s"([^"]+).*
            #  Raised key value 
            keys:
            - client_ip
            - remote_user
            - request_time
            - request_method
            - request_url
            - http_protocol
            - status_code
            - body_bytes_sent
            - handle_time
            - http_referer
            - http_user_agent
            - real_ip
            - log_time
            #  Regular expressions , Will be based on () The capture group extracts the corresponding value
            logRegex: (\S+)\s-[^-]+(\S+)\s\[([^\]]+)\]\s"(\w+)\s(\S+)\s([^"]+)"\s(\d+)\s(\d+)\s(\S+)[^-]+([^"]+)"\s"([^"]+)[^-]+([^"]+)"\s"([^"]+).*
        #  Log collection format ： Fully regular extraction pattern 
        logType: fullregex_log
        topicId: 4511efaa-2cad-4c3c-8c7c-xxxxxxxxx
    inputDetail:
        containerStdout:
            # pod  label 
            includeLabels:
                k8s-app: tke-cls-nginx-std
            namespace: cls-wk
        #  Collection type 
        type: container_stdout

After applying the configuration , The data format of the log service is as follows ：

Container file

• collection deployment by nginx-logfile, The container name is nginx-logfile Log file , The log file path is /data/log/nginx,*.log Final document .

apiVersion: cls.cloud.tencent.com/v1
kind: LogConfig
metadata:
  #  Configuration name , Cluster unique 
  name: www-nginx
spec:
  clsDetail:
    #  Log collection format 
    logType: minimalist_log
    topicId: 56a32d6f-77da-4c4a-a7d1-xxxxxxxxx
  inputDetail:
    containerFile:
      container: nginx-logfile
      #  file name , Support for wildcards * and ？
      filePattern: '*.log'
      logPath: /data/log/nginx
      namespace: cls-wk
      workload:
        kind: deployment
        name: nginx-logfile
    #  The collection type is container file 
    type: container_file

• Separator mode log file

apiVersion: cls.cloud.tencent.com/v1
kind: LogConfig
metadata:
  #  Configuration name , Cluster unique 
  name: cls-delimiter
spec:
  clsDetail:
      #  Separator 
      delimiter: '::'
      #  Extracted key
      keys:
      - remote_addr
      - remote_user
      - time_local
      - request
      - status
      - body_bytes_sent
      - request_time
      - http_referer
      - http_user_agent
      - http_x_forwarded_for
      - msec
      #  The log time uses the field time in the log 
      timeFormat: '%d/%b/%Y %H:%M:%S'
      timeKey: time_local
    #  The format of log collection is separator mode 
    logType: delimiter_log
    topicId: e850733e-f32b-4a8e-a246-xxxxxxxx
  inputDetail:
    containerFile:
      container: nginxdelimiter
      filePattern: '*.log'
      logPath: /data/log/nginx
      namespace: cls-wk
      workload:
        kind: deployment
        name: nginxdelimiter
    #  The collection type is container file 
    type: container_file

After applying the configuration , The data format of the log service is as follows ：

more CLS and TKE Please refer to the following links for log collection ：

CLS The log service

Use CRD Configure log collection

A friendly reminder ：

• adopt CRD Collection configuration TKE journal , The index needs to be opened and configured in the log console ;
• Recommended by TKE Console configuration , By default, full-text index and partial key value index are opened .