配置ck使用zipkin + opentelemetry 辅助函数栈耗时可视化

zipkin可以将ts中的opentelemetry埋点发出的信息进行可视化（信息的转发依赖于物化视图的url外部存储功能）

数据流

流程

代码埋点

在一个作用域中声明span实体，在span的构造析构时会对时间信息进行打点。

在tsearch-client中进行表建立物化视图

参考：https://clickhouse.com/docs/en/operations/opentelemetry

不需要修改，除非zipkin的服务不在本地，端口不在9411.

CREATE MATERIALIZED VIEW default.zipkin_spans
(
    `traceId` String,
    `parentId` String,
    `id` String,
    `name` LowCardinality(String),
    `timestamp` UInt64,
    `duration` Int64,
    `localEndpoint` Tuple(serviceName String),
    `serviceName` String,
    `tags` Map(LowCardinality(String), String)
)
ENGINE = URL('http://127.0.0.1:9411/api/v2/spans', 'JSONEachRow')
SETTINGS output_format_json_named_tuples_as_objects = 1, output_format_json_array_of_rows = 1 AS
SELECT
    lower(hex(trace_id)) AS traceId,
    multiIf(parent_span_id = 0, '0', lower(hex(parent_span_id))) AS parentId,
    lower(hex(span_id)) AS id,
    operation_name AS name,
    start_time_us AS timestamp,
    finish_time_us - start_time_us AS duration,
    CAST(tuple('clickhouse'), 'Tuple(serviceName text)') AS localEndpoint,
    'ser' AS serviceName,
    attribute AS tags
FROM system.opentelemetry_span_log

zipkin

下载 并运行jar包
https://zipkin.io/

查看即可
查看http://127.0.0.1:9411/zipkin

其他问题

1. 从远端转发到本地浏览器查看
   https://www.wyr.me/post/723
2. 如果用docker跑ck，物化视图可能发送不到宿主机的zipkin，应该打通网络。https://wiki.eryajf.net/pages/674f53/#%E6%96%B9%E6%A1%88%E4%BA%8C