Kubernetes Controller Manager¶

视图预览¶

Controller Manager 性能指标，包括 Service Controller Rate Limiter Use、 Deployment Controller Rate Limiter Use、 Daemon Controller Rate Limiter Use、 Replicaset Controller Rate Limiter Use、 CPU 、 Memory、 Goroutines 等。

视图暂无

版本支持¶

操作系统支持：Linux

Kubernetes 版本：1.18+

前置条件¶

Kubernetes 集群 <安装 DataKit>

安装配置¶

说明：示例 Kubernetes 版本为 1.22.6，DataKit 版本为 1.2.17，各个不同版本指标可能存在差异。

部署实施¶

指标采集 (必选)¶

1、 ConfigMap 增加 controller-manager.conf 配置

在部署 DataKit 使用的 datakit.yaml 文件中，ConfigMap 资源中增加 controller-manager.conf 。

apiVersion: v1
kind: ConfigMap
metadata:
  name: datakit-conf
  namespace: datakit
data:
  #### controller-manager  ## 下面是新增部分
  controller-manager.conf: |-
    [[inputs.prom]]
      ## Exporter地址或者文件路径（Exporter地址要加上网络协议http或者https）
      ## 文件路径各个操作系统下不同
      ## Windows example: C:\\Users
      ## UNIX-like example: /usr/local/
      urls = ["https://172.16.0.229:10257/metrics"]

      ## 采集器别名
      source = "prom-controller-manager"

      ## 指标类型过滤, 可选值为 counter, gauge, histogram, summary
      # 默认只采集 counter 和 gauge 类型的指标
      # 如果为空，则不进行过滤
      metric_types = ["counter", "gauge"]

      ## 指标名称过滤
      # 支持正则，可以配置多个，即满足其中之一即可
      # 如果为空，则不进行过滤
      #metric_name_filter = [""]

      ## 指标集名称前缀
      # 配置此项，可以给指标集名称添加前缀
      #measurement_prefix = ""

      ## 指标集名称
      # 默认会将指标名称以下划线"_"进行切割，切割后的第一个字段作为指标集名称，剩下字段作为当前指标名称
      # 如果配置measurement_name, 则不进行指标名称的切割
      # 最终的指标集名称会添加上measurement_prefix前缀
      measurement_name = "prom_controller_manager"

      ## 采集间隔 "ns", "us" (or "µs"), "ms", "s", "m", "h"
      interval = "60s"

      ## 过滤tags, 可配置多个tag
      # 匹配的tag将被忽略
      tags_ignore = ["action","build_date","clusterCIDR","code","compiler","completion_mode","git_commit","git_tree_state","git_version","go_version","major","method","minor","name","operation","platform","plugin_name","result","status","storage_class","username","version","volume_mode"]
      metric_name_filter = ["service_controller_rate_limiter_use","deployment_controller_rate_limiter_use","daemon_controller_rate_limiter_use","replicaset_controller_rate_limiter_use","cronjob_controller_rate_limiter_use","job_controller_rate_limiter_use","serviceaccount_controller_rate_limiter_use","endpoint_controller_rate_limiter_use","replication_controller_rate_limiter_use","gc_controller_rate_limiter_use","serviceaccount_tokens_controller_rate_limiter_use","token_cleaner_rate_limiter_use","node_ipam_controller_rate_limiter_use","node_collector_unhealthy_nodes_in_zone","job_controller_job_finished_total","process_resident_memory_bytes","process_cpu_seconds_total","go_goroutines"]

      ## TLS 配置
      tls_open = true
      #tls_ca = ""
      #tls_cert = ""
      #tls_key = ""

      ## 自定义指标集名称
      # 可以将包含前缀prefix的指标归为一类指标集
      # 自定义指标集名称配置优先measurement_name配置项
      #[[inputs.prom.measurements]]
      #  prefix = ""
      #  name = ""

      ## 自定义认证方式，目前仅支持 Bearer Token
      [inputs.prom.auth]
       type = "bearer_token"
      # token = "xxxxxxxx"
       token_file = "/var/run/secrets/kubernetes.io/serviceaccount/token"

      ## 自定义Tags
      [inputs.prom.tags]
        instance = "172.16.0.229:10257"

参数说明

urls：controller-manager metrics 地址
source：采集器别名
metric_types：指标类型过滤
metric_name_filter：指标名称过滤
measurement_prefix：指标集名称前缀
measurement_name：指标集名称
interval：采集间隔
tags_ignore: 忽略的 tag
metric_name_filter: 保留的指标名
tls_open：是否忽略安全验证 (如果是 HTTPS，请设置为 true，并设置相应证书)，此处为 true
tls_ca：ca 证书路径
type：自定义认证方式，controller manager 使用 bearer_token 认证
token_file：认证文件路径
inputs.prom.tags：请参考插件标签

2、挂载 controller-manager.conf

在 datakit.yaml 文件的 volumeMounts 下面增加下面内容：

- mountPath: /usr/local/datakit/conf.d/prom/controller-manager.conf
  name: datakit-conf
  subPath: controller-manager.conf

3、重启 DataKit

kubectl delete -f datakit.yaml
kubectl apply -f datakit.yaml

指标预览

插件标签 (必选）¶

参数说明

该配置为自定义标签，可以填写任意 key-value 值
以下示例配置完成后，controller-manager 指标都会带有如下类似的标签，可以进行快速查询
采集 controller-manager 指标，必填的 key 是 instance，值是 controller-manager 的地址

    ## 自定义Tags
       [inputs.prom.tags]
          instance = "172.16.0.229:10257"

重启 DataKit

kubectl delete -f datakit.yaml
kubectl apply -f datakit.yaml

场景视图¶

<场景 - 新建仪表板 - 模板库 - 系统视图 - Kubernetes Controller Manager 监控视图>

指标集¶

指标	描述	数据类型
node_collector_unhealthy_nodes_in_zone	Gauge measuring number of not Ready Nodes per zones.	int
deployment_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for deployment_controller.	int
daemon_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for daemon_controller	int
endpoint_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for endpoint_controller.	int
cronjob_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for cronjob_controller.	int
job_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for cronjob_controller.	int
node_ipam_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for node_ipam_controller.	int
replicaset_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for replicaset_controller.	int
replication_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for replication_controller.	int
service_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for service_controller.	int
gc_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for gc_controller.	int
serviceaccount_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for serviceaccount_controller.	int
serviceaccount_tokens_controller_rate_limiter_use	A metric measuring the saturation of the rate limiter for serviceaccount_tokens_controller.	int
job_controller_job_finished_total	The number of finished job.	int
token_cleaner_rate_limiter_use	A metric measuring the saturation of the rate limiter for token_cleaner.	int
process_resident_memory_bytes	Resident memory size in bytes.	B
process_cpu_seconds_total	Total user and system CPU time spent in seconds.	float
go_goroutines	Number of goroutines that currently exist.	int

常见问题排查¶

<无数据上报排查>

进一步阅读¶