Pinpoint

DataKit 内置的 Pinpoint Agent 用于接收，运算，分析 Pinpoint Tracing 协议数据。

配置¶

采集器配置¶

主机安装Kubernetes

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

[[inputs.pinpoint]]
  ## Pinpoint service endpoint for
  ## - Span Server
  ## - Agent Server(unimplemented, for service intactness and compatibility)
  ## - Metadata Server(unimplemented, for service intactness and compatibility)
  ## - Profiler Server(unimplemented, for service intactness and compatibility)
  address = "127.0.0.1:9991"

  ## Keep rare tracing resources list switch.
  ## If some resources are rare enough(not presend in 1 hour), those resource will always send
  ## to data center and do not consider samplers and filters.
  # keep_rare_resource = false

  ## delete trace message
  # del_message = true

  ## Ignore tracing resources map like service:[resources...].
  ## The service name is the full service name in current application.
  ## The resource list is regular expressions uses to block resource names.
  ## If you want to block some resources universally under all services, you can set the
  ## service name as "*". Note: double quotes "" cannot be omitted.
  # [inputs.pinpoint.close_resource]
    # service1 = ["resource1", "resource2", ...]
    # service2 = ["resource1", "resource2", ...]
    # "*" = ["close_resource_under_all_services"]
    # ...

  ## Sampler config uses to set global sampling strategy.
  ## sampling_rate used to set global sampling rate.
  # [inputs.pinpoint.sampler]
    # sampling_rate = 1.0

  # [inputs.pinpoint.tags]
    # key1 = "value1"
    # key2 = "value2"
    # ...

  ## 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.pinpoint.storage]
    # path = "./pinpoint_storage"
    # capacity = 5120

DataKit Pinpoint Agent 监听地址配置项为：

# Pinpoint GRPC service endpoint for
# - Span Server
# - Agent Server(unimplemented, for service intactness and compatibility)
# - Metadata Server(unimplemented, for service intactness and compatibility)
# - Profiler Server(unimplemented, for service intactness and compatibility)
address = "127.0.0.1:9991"

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_PINPOINT_ADDRESS

代理 URL

字段类型: String

采集器配置字段: address

示例: 127.0.0.1:9991
ENV_INPUT_PINPOINT_KEEP_RARE_RESOURCE

保持稀有跟踪资源列表

字段类型: Boolean

采集器配置字段: keep_rare_resource

默认值: false
ENV_INPUT_PINPOINT_DEL_MESSAGE

删除 trace 消息

字段类型: Boolean

采集器配置字段: del_message

默认值: false
ENV_INPUT_PINPOINT_CLOSE_RESOURCE

忽略指定服务器的 tracing（正则匹配）

字段类型: JSON

采集器配置字段: close_resource

示例: {"service1":["resource1","other"],"service2":["resource2","other"]}
ENV_INPUT_PINPOINT_SAMPLER

全局采样率

字段类型: Float

采集器配置字段: sampler

示例: 0.3
ENV_INPUT_PINPOINT_STORAGE

本地缓存路径和大小（MB）

字段类型: JSON

采集器配置字段: storage

示例: {"storage":"./pinpoint_storage", "capacity": 5120}
ENV_INPUT_PINPOINT_TAGS

自定义标签。如果配置文件有同名标签，将会覆盖它

字段类型: JSON

采集器配置字段: tags

示例: {"k1":"v1", "k2":"v2", "k3":"v3"} |

DataKit 中的 Pinpoint Agent 存在以下限制

目前只支持 gRPC 协议
多服务（Agent/Metadata/Stat/Span）合一的服务使用同一个端口
Pinpoint 链路与 DataKit 链路存在差异，详见下文

Pinpoint Agent 配置¶

下载所需的 Pinpoint APM Agent

Pinpoint 支持实现了多语言的 APM Collector 本文档使用 JAVA Agent 进行配置。下载 JAVA APM Collector。

配置 Pinpoint APM Collector，打开 /path_to_pinpoint_agent/pinpoint-root.config 配置相应的多服务端口
- 配置 profiler.transport.module = GRPC
- 配置 profiler.transport.grpc.agent.collector.port = 9991 （即 DataKit Pinpoint Agent 中配置的端口）
- 配置 profiler.transport.grpc.metadata.collector.port = 9991（即 DataKit Pinpoint Agent 中配置的端口）
- 配置 profiler.transport.grpc.stat.collector.port = 9991 （即 DataKit Pinpoint Agent 中配置的端口）
- 配置 profiler.transport.grpc.span.collector.port = 9991 （即 DataKit Pinpoint Agent 中配置的端口）
启动 Pinpoint APM Agent 启动命令

$ java -javaagent:/path_to_pinpoint/pinpoint-bootstrap.jar \
    -Dpinpoint.agentId=agent-id \
    -Dpinpoint.applicationName=app-name \
    -Dpinpoint.config=/path_to_pinpoint/pinpoint-root.config \
    -jar /path_to_your_app.jar

DataKit 链路数据遵循 OpenTracing 协议，DataKit 中一条链路是通过简单的父子（子 span 中存放父 span 的 id）结构串联起来且每个 span 对应一次函数调用

Pinpoint APM 链路数据较为复杂：

父 span 负责产生子 span 的 ID
子 span 中也要存放父 span 的 ID
使用 span event 替代 OpenTracing 中的 span
一个 span 为一个服务的一次应答过程

PinPointV2¶

DataKit 1.19.0 版本重新优化后更改 source 为 PinPointV2。新版本的链路数据重新梳理 SpanChunk 和 Span 的关系、Event 和 Span 的关系、Span 与 Span 的关系。以及 Event 中 startElapsed 和 endElapsed 时间对齐问题。

主要的逻辑点：

缓存 serviceType 服务表，并写到文件中，防止 DataKit 重启而丢失数据。
Span 中的 parentSpanId 不为 -1，则缓存。如 parentSpanId:-1，则根据 spanEvent 中的 nextSpanId 从缓存中取出 Span 拼接到一个链路中。
缓存所有 SpanChunk 中的 event，直到接收到主 Span 才从缓存中全部取出，追加到链路中。
按顺序累加当前 Event 中 startElapsed 作为下一个 Event 的起始时间。
按照 Depth 字段判断当前 Event 的父子级关系。
遇到数据库查询会将 sql 语句替换当前的 '资源' 名称。

链路字段¶

`pinpoint`¶

This is the field description for the trace.

标签

Tag	Description
base_service	Span Base service name
container_host	Container hostname. Available in OpenTelemetry. Optional.
db_host	DB host name: ip or domain name. Optional.
db_name	Database name. Optional.
db_system	Database system name:mysql,oracle... Optional.
dk_fingerprint	DataKit fingerprint is DataKit hostname
endpoint	Endpoint info. Available in SkyWalking, Zipkin. Optional.
env	Application environment info. Available in Jaeger. Optional.
host	Hostname.
http_method	HTTP request method name. Available in DDTrace, OpenTelemetry. Optional.
http_route	HTTP route. Optional.
http_status_code	HTTP response code. Available in DDTrace, OpenTelemetry. Optional.
http_url	HTTP URL. Optional.
operation	Span name
out_host	This is the database host, equivalent to db_host,only DDTrace-go. Optional.
project	Project name. Available in Jaeger. Optional.
service	Service name. Optional.
source_type	Tracing source type
span_type	Span type
status	Span status
version	Application version info. Available in Jaeger. Optional.

指标列表

Metric	Description
duration	Duration of span Type: int Unit: time,μs
message	Origin content of span Type: string Unit: N/A
parent_id	Parent span ID of current span Type: string Unit: N/A
resource	Resource name produce current span Type: string Unit: N/A
span_id	Span id Type: string Unit: N/A
start	start time of span. Type: int Unit: timeStamp,usec
trace_id	Trace id Type: string Unit: N/A

指标字段¶

`pinpoint-metric`¶

标签

Tag	Description
agentVersion	Pinpoint agent version
agent_id	Agent ID
container	Whether it is a container
hostname	Host name
ip	Agent IP
pid	Process ID
ports	Open ports

指标列表

Metric	Description
GcNewCount	Jvm Gc NewCount Type: int Unit: count
GcNewTime	Jvm Gc NewTime Type: int Unit: timeStamp,msec
JvmCpuLoad	Jvm CPU load Type: int Unit: percent,percent
JvmGcOldCount	Jvm Gc Old Count Type: int Unit: count
JvmGcOldTime	Jvm Gc Old Time Type: int Unit: timeStamp,msec
JvmMemoryHeapMax	Jvm Memory Heap Max Type: int Unit: digital,B
JvmMemoryHeapUsed	Jvm Memory Heap Used Type: int Unit: digital,B
JvmMemoryNonHeapMax	Jvm Memory NonHeap Max Type: int Unit: digital,B
JvmMemoryNonHeapUsed	Jvm Memory NonHeap Used Type: int Unit: digital,B
PoolCodeCacheUsed	Jvm Pool Code Cache Used Type: float Unit: digital,B
PoolMetaspaceUsed	Jvm Pool meta space used Type: float Unit: count
PoolNewGenUsed	Jvm Pool New GenUsed Type: float Unit: digital,B
PoolOldGenUsed	Duration of Jvm garbage collection actions Type: float Unit: digital,B
PoolPermGenUsed	The maximum file descriptor count Type: float Unit: count
PoolSurvivorSpaceUsed	Jvm Pool Survivor SpaceUsed Type: float Unit: digital,B
SystemCpuLoad	system CPU load Type: int Unit: percent,percent

Pinpoint

配置¶

采集器配置¶

Pinpoint Agent 配置¶

PinPointV2¶

链路字段¶

`pinpoint`¶

指标字段¶

`pinpoint-metric`¶

Pinpoint 参考资料¶

文档内容是否对您有帮助？ ×

Pinpoint

配置¶

采集器配置¶

Pinpoint Agent 配置¶

PinPointV2¶

链路字段¶

pinpoint¶

指标字段¶

pinpoint-metric¶

Pinpoint 参考资料¶

文档内容是否对您有帮助？ ×

`pinpoint`¶

`pinpoint-metric`¶