Observability

Sentry
Transactions vs Spans
- from https://docs.sentry.io/product/sentry-basics/tracing/distributed-tracing/
- Span
  - Spans in OpenTelemetry
  - A Span represents a unit of work or operation. Spans are the building blocks of Traces.
  - A Span is best used to when the operation has a start and an end.
  - A Span Event is a information of the Span. It’s best used to track operations with singular point in time.

LLM Link to heading

arize
baserun - From identifying an issue to evaluating the solution
Datadog - OpenAI Monitoring
fiddler -
Gantry - observability, analytics, and evaluation for your AI-powered products.
Helicone - monitoring, logging, and tracing for your LLM applications out of the box #OpenSource
OpenInference - set of conventions and plugins that is complimentary to OpenTelemetry #OpenSource
OpenLLMetry - set of extensions built on top of OpenTelemetry #OpenSource
Phoenix - Evaluate, troubleshoot, and fine tune your LLM, CV, and NLP models in a notebook #OpenSource
Vellum - prompt engineering, semantic search, version control, quantitative testing, and performance monitoring.

OpenTelemetry Link to heading

Python SDK
Python instrumentation - API doc - GitHub - opentelemetry-python-contrib

How to set the trace ID manually? - Discussion - Docs

  tracer = get_tracer()
  propagator = get_global_textmap()
  headers = request.headers
  context = propagator.extract(headers)

  with tracer.start_as_current_span("span_name", context=context):
      ...
  ```

How to rename RequestsInstrumentor events? passing a function for the name_callback argument

How to manually add trace_id to a transaction

  context = context=get_global_textmap().extract(app.current_request.headers)
  with tracer.start_as_current_span("span_name", content=content):
    ...
  ```
- [Example](https://github.com/getsentry/sentry-python/blob/1e3e1097e104abb39799b59654bf4f8725448909/sentry_sdk/integrations/opentelemetry/propagator.py#L45) of a propagator (textmap) extract