AgentBase AgentOps LangChain
agentbase-agentops-langchain 为用户现有的 LangChain Agent 增加 AppBase AgentOps
可观测能力。挂载一次 AgentOpsMiddleware 后,SDK 自动采集 Agent、Chain、LLM、Tool
和 Retriever 生命周期,并通过 OTLP/HTTP 发送到 AppBase AgentOps。
SDK 不创建或替换模型,不代理模型请求,也不改变应用现有的模型 Provider、模型凭据、 Prompt、Tool 或业务逻辑。
安装
python -m pip install agentbase-agentops-langchain
配置
export AGENTBASE_AGENTOPS_ENDPOINT="https://<appbase-host>"
export AGENTBASE_PROJECT_ID="<project-id>"
export AGENTBASE_API_KEY="<project-api-key>"
Project API Key 至少需要 observability.ingest.write 权限。SDK 会自动将 AppBase host 或
/v1 base URL 补齐为完整 OTLP Trace endpoint。
接入现有 Agent
from agentbase_agentops_langchain import AgentOpsMiddleware
from langchain.agents import create_agent
agentops = AgentOpsMiddleware()
agent = create_agent(
model=model,
tools=tools,
middleware=[agentops],
)
result = agent.invoke(
inputs,
config={
"metadata": {
"session_id": conversation_id,
"user_id": authenticated_user_id,
},
"tags": [environment, agent_name, agent_version],
},
)
如果应用已经配置其他 LangChain middleware,请将 AgentOpsMiddleware 追加到原列表,
不要覆盖现有 middleware。后续 invoke、ainvoke、stream、astream、batch
和 abatch 调用不需要手工传入 AgentOps callback。
短生命周期进程
CLI、Job 和测试脚本应在退出前主动 flush:
agentops.flush(timeout_millis=10_000)
flush() 表示同步尝试导出本地队列,不代表 AppBase 服务端已经完成持久化。需要确认接入
成功时,应通过 Console 或 Trace API 等待完整的 Agent、Generation 和 Tool 观测树。
内容采集
默认 AGENTBASE_AGENTOPS_CAPTURE_CONTENT=false,不会上传 Prompt、Response、Tool 参数
或 Retriever 文档正文。只有显式开启后才采集正文:
export AGENTBASE_AGENTOPS_CAPTURE_CONTENT=true
生产环境开启前应确认内容符合数据安全和脱敏要求。
NeMo Relay
已经使用 NVIDIA NeMo Relay 的应用不应再挂载 AgentOpsMiddleware。保留
NemoRelayMiddleware、NemoRelayCallbackHandler 和 Relay scope,将 Relay 的
OTLP/HTTP Trace endpoint 与鉴权 header 指向 AppBase AgentOps。
Session/User 关联需要同时满足:
plugins.toml配置promote_metadata_prefixes = ["langfuse."];- Relay scope metadata 传入
langfuse.session.id和langfuse.user.id。
本页提供可直接完成安装和接入的最小配置。完整的配置、验证、隐私边界和排障说明, 请以 AppBase 官方客户文档为准。
兼容范围
- Python
>=3.10 - LangChain
>=1.0,<2.0 - OTLP/HTTP Trace
当前不提供 OTLP/gRPC、Metrics 或 OpenTelemetry Logs 接入。
Metadata
Release files for agentbase-agentops-langchain 0.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| agentbase_agentops_langchain-0.1.0.tar.gz | 27.2 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| agentbase_agentops_langchain-0.1.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 48.0 kB
Release files / agentbase_agentops_langchain-0.1.0.tar.gz
| Download URL | agentbase_agentops_langchain-0.1.0.tar.gz |
|---|---|
| Size | 27.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
54cd444ce7483aabe9f7cfc42e2f3ddd70730b5adab7fd2373800c55692db1e3
|
|
BLAKE2b-256 checksum How to use checksums |
b7a5bac374053e455a44ec9f501c772cfbd5df04e58443679040710821c5832b
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.13.9
|
Release files / agentbase_agentops_langchain-0.1.0-py3-none-any.whl
| Download URL | agentbase_agentops_langchain-0.1.0-py3-none-any.whl |
|---|---|
| Size | 20.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
7160fd3a4e931c5657c8cda08ea4fa845b16456fe4445bf7bb466a5624e43895
|
|
BLAKE2b-256 checksum How to use checksums |
725ec61a05d676e820ee70c21b658251da35cddfaaee37bbdaf84823244832cd
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.13.9
|
Security for agentbase-agentops-langchain 0.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.