返回目录

dsh-otel

精选清单收录维护状态: 活跃

tma1-ai/dsh-otel

为 DeepSeek Harness 接入 OpenTelemetry 的链路、指标与日志,直接写入 GreptimeDB。

前往 GitHub项目主页
$ dsh plugin add dsh-otel

安装

dsh 没有统一的安装命令——把该插件 README 中的配置行加入你的 profile/patch 配置,然后重启即可。

了解安装方式

10

星标

3

Fork

TypeScript

语言

Apache-2.0

许可证

2026-08-25

创建于

2026-09-03

最近推送

README

@tma1-ai/dsh-plugin-greptimedb

npm CI node license

English | 中文

What a DeepSeek Harness run costs you in tokens, money, and time, written into GreptimeDB as OpenTelemetry traces, metrics, and logs.

Seven Grafana dashboards read it back. No collector. No sidecar. No fork of DSH. It installs as an ordinary plugin, and every turn, model call, and tool execution becomes a row you can query:

-- Slowest tool calls, with the model that requested them.
SELECT span_name, model, duration_nano / 1000000 AS ms
FROM (
  SELECT span_name, duration_nano,
         MAX("span_attributes.gen_ai.request.model")
           OVER (PARTITION BY trace_id, "span_attributes.dsh.step") AS model
  FROM opentelemetry_traces
  WHERE "span_attributes.dsh.step" IS NOT NULL
)
WHERE span_name LIKE 'execute_tool%'
ORDER BY duration_nano DESC
LIMIT 10;

Overview

Quick start

Requires pnpm 10 or newer. dsh plugin forwards to whichever pnpm is on your PATH, and a dsh profile directory is its own pnpm workspace root. pnpm 9 refuses to install there and ignores the linker settings dsh writes.

Start the database and Grafana. The compose stack under grafana/ brings up GreptimeDB with the seven dashboards provisioned. Grafana reads those dashboards off disk, so fetch that one directory instead of cloning the repository:

curl -fsSL https://github.com/tma1-ai/dsh-otel/archive/main.tar.gz \
  | tar -xz --strip-components=1 dsh-otel-main/grafana
cd grafana && docker compose up -d

Install the plugin. Its defaults point at the database started above, so no configuration is required:

dsh plugin --profile web add @tma1-ai/dsh-plugin-greptimedb

The package ships a bundle patch, so that one command wires it into the profile.

Run DSH and check the data. Traces and logs are written within scheduledDelayMillis, 5 seconds by default; metrics on the next collection period, 30 seconds by default. DSH produces no data while idle, so run a task first:

dsh web

Grafana is at http://localhost:3000, with anonymous admin access enabled and no login required. Start with the Overview dashboard.

For the database alone, GreptimeDB's own console at http://localhost:4000/dashboard/ is enough to check the tables and run ad-hoc SQL:

docker run -p 127.0.0.1:4000-4003:4000-4003 \
  -v "$(pwd)/greptimedb_data:/greptimedb_data" \
  --name greptime --rm greptime/greptimedb:v1.2.0-beta.2 standalone start \
  --http-addr 0.0.0.0:4000 --rpc-bind-addr 0.0.0.0:4001 \
  --mysql-addr 0.0.0.0:4002 --postgres-addr 0.0.0.0:4003

Both bind port 4000, so run one or the other.

Compatibility

DSH package APIs change between prereleases, so each plugin release targets one DSH version:

Plugin npm tag DSH
0.1.0-beta.5 beta 0.1.2-alpha.5
0.1.0-beta.4 and earlier latest 0.1.1-rc.2

dsh plugin add with no version installs latest, which tracks DSH's own latest. Ask for the beta tag to follow the DSH alpha line — arguments are forwarded to pnpm add verbatim, so an exact version works the same way:

dsh plugin --profile web add @tma1-ai/dsh-plugin-greptimedb@beta

A mismatch is not a load failure. The plugin mounts, then drops every record with a greptimedb-otel: dropped a telemetry record warning, and the tables stay empty.

Configuration

To point the plugin at your own database, override the row in $DSH_HOME/profiles/<name>/cordis.patch.yml:

- id: greptimedb-otel
  name: '@tma1-ai/dsh-plugin-greptimedb'
  config:
    endpoint: https://<host>/v1/otlp
    database: <dbname>
    username: <user>
    password: <password>

A profile patch replaces the row's whole config instead of merging into it, so restate every field you want to keep.

Key Default Notes
endpoint (required) OTLP base URL, e.g. http://localhost:4000/v1/otlp. The plugin appends each signal's /v1/{traces,metrics,logs} suffix; a per-signal path is rejected at load.
database public Sent as X-Greptime-DB-Name.
username / password (none) Basic auth. Both or neither.
signals all three Any subset of traces, metrics, logs. A disabled signal builds no exporter.
content none How much payload may leave the process. See What leaves the machine.
serviceName dsh OTel service.name.
logTable / traceTable GreptimeDB defaults Destination table overrides.
ttl 180d Retention for the log and trace tables this plugin creates, sent as x-greptime-hints. Also accepts forever. GreptimeDB applies it when it auto-creates the table; an existing table keeps its own until ALTER TABLE. Metric tables are not covered — see Known limitations. Set it empty to send no hint and inherit the database default.
shutdownTimeoutMillis 3000 Deadline for the entire teardown sequence.
metricIntervalMillis 30000 Metric collection period. Must be at least exportTimeoutMillis.
maxExportBatchSize / maxQueueSize 512 / 2048 Batch and buffer bounds.
scheduledDelayMillis / exportTimeoutMillis 5000 / 30000 Export cadence and per-request deadline.

在 GitHub 阅读完整 README →

DSH Plugins 是独立的 DeepSeek Harness 插件目录,与 DeepSeek 官方无关,也不代表官方背书。第三方插件未经安全审计,安装前请审查源码。

每周获取最新的 DeepSeek Harness 插件,绝不滥发。