跳转至

网络拨测

该采集器是网络拨测结果数据采集，所有拨测产生的数据，上报观测云。

配置

私有拨测节点部署，需在 观测云页面创建私有拨测节点。创建完成后，将页面上相关信息填入 conf.d/network/dialtesting.conf 即可：

进入 DataKit 安装目录下的 conf.d/network 目录，复制 dialtesting.conf.sample 并命名为 dialtesting.conf。示例如下：

[[inputs.dialtesting]]
  # We can also configure a JSON path like "file:///your/dir/json-file-name"
  server = "https://dflux-dial.guance.com"

  # [require] node ID
  region_id = "default"

  # if server are dflux-dial.guance.com, ak/sk required
  ak = ""
  sk = ""

  # The interval to pull the tasks.
  pull_interval = "1m"

  # The timeout for the HTTP request.
  time_out = "30s"

  # The number of the workers.
  workers = 6

  # Collect related metric when job execution time error interval is larger than task_exec_time_interval
  task_exec_time_interval = "5s"

  # Stop the task when the task failed to send data to dataway over max_send_fail_count.
  max_send_fail_count = 16

  # The max sleep time when send data to dataway failed.
  max_send_fail_sleep_time = "30m"

  # The max number of jobs sending data to dataway in parallel. Default 10.
  max_job_number = 10

  # The max number of job chan. Default 1000.
  max_job_chan_number = 1000

  # Disable internal network task.
  disable_internal_network_task = true

  # Disable internal network cidr list.
  disabled_internal_network_cidr_list = []

  # Custom tags.
  [inputs.dialtesting.tags]
  # some_tag = "some_value"
  # more_tag = "some_other_value"
  # ...

配置好后，重启 DataKit 即可。

可通过 ConfigMap 方式注入采集器配置 或 配置 ENV_DATAKIT_INPUTS 开启采集器。

也支持以环境变量的方式修改配置参数（需要在 ENV_DEFAULT_ENABLED_INPUTS 中加为默认采集器）：

  • ENV_INPUT_DIALTESTING_ENV_INPUT_DIALTESTING_DISABLE_INTERNAL_NETWORK_TASK

    是否允许内网地址/服务的拨测。默认不允许

    字段类型: Boolean

    采集器配置字段: disable_internal_network_task

    示例: true

    默认值: false

  • ENV_INPUT_DIALTESTING_ENV_INPUT_DIALTESTING_DISABLED_INTERNAL_NETWORK_CIDR_LIST

    禁止拨测的 CIDR 地址列表

    字段类型: List

    采集器配置字段: disabled_internal_network_cidr_list

    示例: ["192.168.0.0/16"]

    默认值: -

  • ENV_INPUT_DIALTESTING_ENV_INPUT_DIALTESTING_ENABLE_DEBUG_API

    禁止拨测调试接口（默认禁止）

    字段类型: Boolean

    采集器配置字段: env_input_dialtesting_enable_debug_api

    示例: false

    默认值: false

Note

目前只有 Linux 的拨测节点才支持「路由跟踪」，跟踪数据会保存在相关指标的 traceroute 字段中。

拨测节点部署

以下是拨测节点的网络部署拓扑图，这里存在两种拨测节点部署方式：

  • 公网拨测节点：直接使用观测云在全球部署的拨测节点来检测 公网 的服务运行情况。
  • 私网拨测节点：如果需要拨测用户 内网 的服务，此时需要用户自行部署 私有 的拨测节点。当让，如果网络允许，这些私有的拨测节点也能部署公网上的服务。

不管是公网拨测节点，还是私有拨测节点，它们都能通过 Web 页面创建拨测任务。

graph TD
  %% node definitions
  dt_web(拨测 Web UI);
  dt_db(拨测任务公网存储);
  dt_pub(DataKit 公网拨测节点);
  dt_pri(DataKit 私有拨测节点);
  site_inner(内网站点);
  site_pub(公网站点);
  dw_inner(内网 Dataway);
  dw_pub(公网 Dataway);
  server(观测云);

  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

  dt_web -->|创建拨测任务| dt_db;
  dt_db -->|拉取拨测任务| dt_pub -->|拨测结果| dw_pub --> server;
  dt_db -->|拉取拨测任务| dt_pri;
  dt_pub <-->|实施拨测| site_pub;

  dt_pri <-.->|实施拨测| site_pub;
  dw_inner --> server;
  subgraph "用户内网"
  dt_pri <-->|实施拨测| site_inner;
  dt_pri -->|拨测结果| dw_inner;
  end

日志

http_dial_testing

  • 标签
Tag Description
city The name of the city
country The name of the country
datakit_version The DataKit version
dest_ip The IP address of the destination
df_label The label of the task
internal The boolean value, true for domestic and false for overseas
isp ISP, such as chinamobile, chinaunicom, chinatelecom
method HTTP method, such as GET
name The name of the task
node_name The name of the node
owner The owner name
proto The protocol of the HTTP, such as 'HTTP/1.1'
province The name of the province
status The status of the task, either 'OK' or 'FAIL'
status_code_class The class of the status code, such as '2xx'
status_code_string The status string, such as '200 OK'
url The URL of the endpoint to be monitored
  • 指标列表
Metric Description
config_vars The configuration variables of the task
Type: string
Unit: -
fail_reason The reason that leads to the failure of the task
Type: string
Unit: -
message The message string which includes the header and the body of the request or the response
Type: string
Unit: -
response_body_size The length of the body of the response
Type: int
Unit: digital,B
response_connection HTTP connection time
Type: float
Unit: time,μs
response_dns HTTP DNS parsing time
Type: float
Unit: time,μs
response_download HTTP downloading time
Type: float
Unit: time,μs
response_ssl HTTP ssl handshake time
Type: float
Unit: time,μs
response_time The time of the response
Type: int
Unit: time,μs
response_ttfb HTTP response ttfb
Type: float
Unit: time,μs
seq_number The sequence number of the test
Type: int
Unit: count
status_code The response code
Type: int
Unit: -
success The number to specify whether is successful, 1 for success, -1 for failure
Type: int
Unit: -
task The raw task string
Type: string
Unit: -

tcp_dial_testing

  • 标签
Tag Description
city The name of the city
country The name of the country
datakit_version The DataKit version
dest_host The name of the host to be monitored
dest_ip The IP address
dest_port The port of the TCP connection
df_label The label of the task
internal The boolean value, true for domestic and false for overseas
isp ISP, such as chinamobile, chinaunicom, chinatelecom
name The name of the task
node_name The name of the node
owner The owner name
proto The protocol of the task
province The name of the province
status The status of the task, either 'OK' or 'FAIL'
  • 指标列表
Metric Description
config_vars The configuration variables of the task
Type: string
Unit: -
fail_reason The reason that leads to the failure of the task
Type: string
Unit: -
message The message string includes the response time or fail reason
Type: string
Unit: -
response_time The time of the response
Type: int
Unit: time,μs
response_time_with_dns The time of the response, which contains DNS time
Type: int
Unit: time,μs
seq_number The sequence number of the test
Type: int
Unit: count
success The number to specify whether is successful, 1 for success, -1 for failure
Type: int
Unit: -
task The raw task string
Type: string
Unit: -
traceroute The json string fo the traceroute result
Type: string
Unit: -

icmp_dial_testing

  • 标签
Tag Description
city The name of the city
country The name of the country
datakit_version The DataKit version
dest_host The name of the host to be monitored
df_label The label of the task
internal The boolean value, true for domestic and false for overseas
isp ISP, such as chinamobile, chinaunicom, chinatelecom
name The name of the task
node_name The name of the node
owner The owner name
proto The protocol of the task
province The name of the province
status The status of the task, either 'OK' or 'FAIL'
  • 指标列表
Metric Description
average_round_trip_time The average time of the round trip(RTT)
Type: float
Unit: time,μs
average_round_trip_time_in_millis The average time of the round trip(RTT), deprecated
Type: float
Unit: time,ms
config_vars The configuration variables of the task
Type: string
Unit: -
fail_reason The reason that leads to the failure of the task
Type: string
Unit: -
max_round_trip_time The maximum time of the round trip(RTT)
Type: float
Unit: time,μs
max_round_trip_time_in_millis The maximum time of the round trip(RTT), deprecated
Type: float
Unit: time,ms
message The message string includes the average time of the round trip or the failure reason
Type: string
Unit: -
min_round_trip_time The minimum time of the round trip(RTT)
Type: float
Unit: time,μs
min_round_trip_time_in_millis The minimum time of the round trip(RTT), deprecated
Type: float
Unit: time,ms
packet_loss_percent The loss percent of the packets
Type: float
Unit: -
packets_received The number of the packets received
Type: int
Unit: count
packets_sent The number of the packets sent
Type: int
Unit: count
seq_number The sequence number of the test
Type: int
Unit: count
std_round_trip_time The standard deviation of the round trip
Type: float
Unit: time,μs
std_round_trip_time_in_millis The standard deviation of the round trip, deprecated
Type: float
Unit: time,ms
success The number to specify whether is successful, 1 for success, -1 for failure
Type: int
Unit: -
task The raw task string
Type: string
Unit: -
traceroute The json string fo the traceroute result
Type: string
Unit: -

websocket_dial_testing

  • 标签
Tag Description
city The name of the city
country The name of the country
datakit_version The DataKit version
df_label The label of the task
internal The boolean value, true for domestic and false for overseas
isp ISP, such as chinamobile, chinaunicom, chinatelecom
name The name of the task
node_name The name of the node
owner The owner name
proto The protocol of the task
province The name of the province
status The status of the task, either 'OK' or 'FAIL'
url The URL string, such as ws://www.abc.com
  • 指标列表
Metric Description
config_vars The configuration variables of the task
Type: string
Unit: -
fail_reason The reason that leads to the failure of the task
Type: string
Unit: -
message The message string includes the response time or the failure reason
Type: string
Unit: -
response_message The message of the response
Type: string
Unit: -
response_time The time of the response
Type: int
Unit: time,μs
response_time_with_dns The time of the response, include DNS
Type: int
Unit: time,μs
sent_message The sent message
Type: string
Unit: -
seq_number The sequence number of the test
Type: int
Unit: count
success The number to specify whether is successful, 1 for success, -1 for failure
Type: int
Unit: -
task The raw task string
Type: string
Unit: -

multi_dial_testing

  • 标签
Tag Description
city The name of the city
country The name of the country
datakit_version The DataKit version
df_label The label of the task
internal The boolean value, true for domestic and false for overseas
isp ISP, such as chinamobile, chinaunicom, chinatelecom
name The name of the task
node_name The name of the node
owner The owner name
province The name of the province
status The status of the task, either 'OK' or 'FAIL'
  • 指标列表
Metric Description
config_vars The configuration variables of the task
Type: string
Unit: -
fail_reason The reason that leads to the failure of the task
Type: string
Unit: -
last_step The last number of the task be executed
Type: int
Unit: -
message The message string which includes the header and the body of the request or the response
Type: string
Unit: -
response_time The time of the response
Type: int
Unit: time,μs
seq_number The sequence number of the test
Type: int
Unit: count
steps The result of each step
Type: string
Unit: -
success The number to specify whether is successful, 1 for success, -1 for failure
Type: int
Unit: -
task The raw task string
Type: string
Unit: -

traceroute

traceroute 是「路由跟踪」数据的 JSON 文本，整个数据是一个数组对象，对象中的每个数组元素记录了一次路由探测的相关情况，示例如下：

[
    {
        "total": 2,
        "failed": 0,
        "loss": 0,
        "avg_cost": 12700395,
        "min_cost": 11902041,
        "max_cost": 13498750,
        "std_cost": 1129043,
        "items": [
            {
                "ip": "10.8.9.1",
                "response_time": 13498750
            },
            {
                "ip": "10.8.9.1",
                "response_time": 11902041
            }
        ]
    },
    {
        "total": 2,
        "failed": 0,
        "loss": 0,
        "avg_cost": 13775021,
        "min_cost": 13740084,
        "max_cost": 13809959,
        "std_cost": 49409,
        "items": [
            {
                "ip": "10.12.168.218",
                "response_time": 13740084
            },
            {
                "ip": "10.12.168.218",
                "response_time": 13809959
            }
        ]
    }
]

字段描述：

字段 类型 说明
total number 总探测次数
failed number 失败次数
loss number 失败百分比
avg_cost number 平均耗时(μs)
min_cost number 最小耗时(μs)
max_cost number 最大耗时(μs)
std_cost number 耗时标准差(μs)
items Item 的 Array 每次探测信息(详见下面 items 字段说明)

items 字段说明

字段 类型 说明
ip string IP 地址，如果失败，值为 *
response_time number 响应时间(μs)

拨测采集器自身指标采集

拨测采集器会暴露 Prometheus 指标，如果需要上报这些指标至观测云，可以通过 DataKit 采集器 进行采集，相关配置参考如下：

[[inputs.dk]]
  ......

  metric_name_filter = [

  ### others...

  ### dialtesting
  "datakit_dialtesting_.*",

  ]

  ......

文档评价

文档内容是否对您有帮助？ ×