在Docker中部署DataKit最佳实践

本文主要介绍如何在 Docker 中安装 DataKit。

配置和启动 DataKit 容器

登陆观测云平台，点击「集成」 -「DataKit」 - 「Docker」，然后拷贝第二步的启动命令，启动参数按实际情况配置。

拷贝启动命令：

sudo docker run \--hostname "$(hostname)" \--workdir /usr/local/datakit \-v "/etc/conf/dir/conf.d":"/usr/local/datakit/conf.d/host-inputs-conf"-v "/":"/rootfs" \-v /var/run/docker.sock:/var/run/docker.sock \-e ENV_DATAWAY="https://openway.guance.com?token=tkn_XXXX" \-e ENV_DEFAULT_ENABLED_INPUTS='cpu,disk,diskio,mem,swap,system,net,host_processes,hostobject,container,dk' \-e ENV_GLOBAL_HOST_TAGS="tag1=a1,tag2=a2" \-e ENV_HTTP_LISTEN="0.0.0.0:9529" \-e HOST_PROC="/rootfs/proc" \-e HOST_SYS="/rootfs/sys" \-e HOST_ETC="/rootfs/etc" \-e HOST_VAR="/rootfs/var" \-e HOST_RUN="/rootfs/run" \-e HOST_DEV="/rootfs/dev" \-e HOST_ROOT="/rootfs" \--cpus 2 \--memory 1g \--privileged \--publish 9529:9529 \--name datakit-docker \-d \pubrepo.guance.com/datakit/datakit:1.66.2

容器启动后，查看是否启动成功：

docker ps

如下所示，启动成功：

启动参数说明：

• --hostname：将宿主机的主机名作为 DataKit 运行的主机名，如果需要在当前宿主机上运行多个 DataKit，可以给它适当加一些后缀 --hostname "$(hostname)-dk1"
• --workdir：设置容器工作目录
• -v：各种宿主机文件挂载：
  • DataKit 中有很多配置文件，我们可以将其在宿主机上准备好，通过 -v 一次性整个挂载到容器中去（容器中的路径为 conf.d/host-inputs-conf 目录）
  • 此处将宿主机根目录挂载进 Datakit，目的是访问宿主机上的各种信息（比如 /proc 目录下的各种文件），便于默认开启的采集器采集数据
  • 将 docker.sock 文件挂载进 Datakit 容器，便于 container 采集器采集数据。不同宿主机该文件目录可能不同，需按照实际来配置
• -e：各种 Datakit 运行期的环境变量配置，这些环境变量功能跟 DaemonSet 部署 时是一样的
• ENV_DATAWAY : 将 token 粘贴到 ENV_DATAWAY 环境变量值中 “token=”
• --publish：便于外部将 Trace 等数据发送给 Datakit 容器，此处我们将 Datakit 的 HTTP 端口映射到外面的 9529 上，诸如 trace 数据设置发送地址的时候，需关注这个端口设置。
• --name: 指定 Docker 容器名称，否则，name 将随机生成
• 此处对该运行的 DataKit 设置了 2C 的 CPU 和 1GiB 内存限制

假如我们在 /host/conf/dir 目录下配置了如下一些采集器：

• APM：DDTrace/OpenTelemetry 等采集器
• Prometheuse exporter：在当前 docker 环境中，某些应用容器暴露了自身指标（一般形如 http://ip:9100/metrics），那么我们可以将其端口暴露出来，然后编写 prom.conf 来采集这些指标
• 日志采集：如果某些 Docker 容器将日志写入了宿主机的某个目录，我们可以单独编写日志采集配置来采集这些文件。不过事先我们需要通过 -v 将这些宿主机的目录挂载进 Datakit 容器。另外，默认开启的 container 采集器，会自动采集所有容器的 stdout 日志

登陆观测云平台，点击「基础设施」 - 「容器」，查看名称为 datakit-docker 容器是否上报，点击进入查看容器详情。

场景演示

如何使用 Docker 的 DataKit 采集用户应用访问数据。

开启 RUM 采集器

在挂载的目录 /etc/conf/dir/conf.d 下创建 rum 目录，然后在 rum 目录下，新建 rum.conf 文件，内容如下：

# {"version": "1.66.2", "desc": "do NOT edit this line"}                                                                                 [[inputs.rum]]                                                                                                                           ## profile Agent endpoints register by version respectively.                                                                           ## Endpoints can be skipped listen by remove them from the list.                                                                       ## Default value set as below. DO NOT MODIFY THESE ENDPOINTS if not necessary.                                                         endpoints = ["/v1/write/rum"]                                                                                                          ## used to upload rum session replay.                                                                                                  session_replay_endpoints = ["/v1/write/rum/replay"]                                                                                    ## specify which metrics should be captured.                                                                                           measurements = ["view", "resource", "action", "long_task", "error", "telemetry"]                                                       ## Android command-line-tools HOME                                                                                                     android_cmdline_home = "/usr/local/datakit/data/rum/tools/cmdline-tools"                                                               ## proguard HOME                                                                                                                       proguard_home = "/usr/local/datakit/data/rum/tools/proguard"                                                                           ## android-ndk HOME                                                                                                                    ndk_home = "/usr/local/datakit/data/rum/tools/android-ndk"                                                                             ## atos or atosl bin path                                                                                                              ## for macOS datakit use the built-in tool atos default                                                                                ## for Linux there are several tools that can be used to instead of macOS atos partially,                                              ## such as https://github.com/everettjf/atosl-rs                                                                                       atos_bin_path = "/usr/local/datakit/data/rum/tools/atosl"                                                                              # Provide a list to resolve CDN of your static resource.                                                                               # Below is the Datakit default built-in CDN list, you can uncomment that and change it to your cdn list,                               # it's a JSON array like: [{"domain": "CDN domain", "name": "CDN human readable name", "website": "CDN official website"},...],        # domain field value can contains '*' as wildcard, for example: "kunlun*.com",                                                         # it will match "kunluna.com", "kunlunab.com" and "kunlunabc.com" but not "kunlunab.c.com".                                            # cdn_map = '''                                                                                                                        # [                                                                                                                                    #   {"domain":"15cdn.com","name":"some-CDN-name","website":"https://www.15cdn.com"},                                                   #   {"domain":"tzcdn.cn","name":"some-CDN-name","website":"https://www.15cdn.com"}                                                     # ]                                                                                                                                    # '''                                                                                                                                  ## Threads config controls how many goroutines an agent cloud start to handle HTTP request.                                            ## buffer is the size of jobs' buffering of worker channel.                                                                            ## threads is the total number fo goroutines at running time.                                                                          # [inputs.rum.threads]                                                                                                                 #   buffer = 100                                                                                                                       #   threads = 8                                                                                                                        ## Storage config a local storage space in hard dirver to cache trace data.                                                            ## path is the local file path used to cache data.                                                                                     ## capacity is total space size(MB) used to store data.                                                                                # [inputs.rum.storage]                                                                                                                 #   path = "./rum_storage"                                                                                                             #   capacity = 5120                                                                                                                    ## session_replay config is used to control Session Replay uploading behavior.                                                         ## cache_path set the disk directory where temporarily cache session replay data.                                                      ## cache_capacity_mb specify the max storage space (in MiB) that session replay cache can use.                                         ## clear_cache_on_start set whether we should clear all previous session replay cache on restarting Datakit.                           ## upload_workers set the count of session replay uploading workers.                                                                   ## send_timeout specify the http timeout when uploading session replay data to dataway.                                                ## send_retry_count set the max retry count when sending every session replay request.                                                 ## filter_rules set the the filtering rules that matched session replay data will be dropped,                                          ## all rules are of relationship OR, that is to day, the data match any one of them will be dropped.                                   # [inputs.rum.session_replay]                                                                                                          #   cache_path = "/usr/local/datakit/cache/session_replay"                                                                             #   cache_capacity_mb = 20480                                                                                                          #   clear_cache_on_start = false                                                                                                       #   upload_workers = 16                                                                                                                #   send_timeout = "75s"                                                                                                               #   send_retry_count = 3                                                                                                               #   filter_rules = [                                                                                                                   #       "{ service = 'xxx' or version IN [ 'v1', 'v2'] }",                                                                             #       "{ app_id = 'yyy' and env = 'production' }"                                                                                    #   ]                                       

然后重启 DataKit。

docker restart datakit-docker
docker ps

进入容器查看是否挂载成功,如下图所示已成功挂载。

docker exec -it datakit-docker /bin/bash
datakit monitor

应用接入

登录观测云控制台，进入「用户访问监测」，点击左上角「新建应用」，即可开始创建一个新的应用。

选择 Web 应用，并选择本地环境部署的 NPM 接入方式。

按需填入配置参数，点击创建，即可在应用列表查看应用。

然后，将 SDK 复制到前端项目中。

启动应用后，进行访问，相关数据会上报到观测云平台。

观测云效果

登录观测云控制台，点击「用户访问监测」 -「应用列表」，然后点击创建的应用。

点击查看器，就能查询采集到的用户访问数据。