palqee-prisma-otel 0.6.2

Prisma OpenTelemetry SDK - Lightweight wrapper for Oversight Intelligence platform

Lightweight OpenTelemetry wrapper that ships traces and spans to the Palqee Prisma oversight intelligence platform.

---

What is this?

palqee-prisma-otel is the transport layer for getting OpenTelemetry traces into Palqee Prisma. It configures an OTLP HTTP exporter pre-wired to authenticate against Prisma's Receiver and tag every span with the resource attributes Prisma uses to route traces to the right run and project.

Most applications should reach for palqee-prisma-ai — the high-level SDK that bundles this wrapper, attaches validators, and exposes a one-line init(). Pick palqee-prisma-otel directly when you need:

• Tracing without the full SDK — sidecars, batch jobs, or libraries that emit spans but never call validators.
• An existing OpenTelemetry setup — you already configure tracers, processors, and propagators yourself, and just need a Prisma-aware exporter.
• Tight control over span emission — long-lived services that manage their own TracerProvider lifecycle.

Installation

pip install palqee-prisma-otel

Imports as palqee_prisma_otel:

from palqee_prisma_otel import configure_tracing

Requires Python 3.10+.

Quickstart

import os
from palqee_prisma_otel import configure_tracing

configure_tracing(
    api_key=os.environ["PALQEE_PRISMA_API_KEY"],
    service_name="complaints-review",
    run_id="run-abc123",         # routes spans to this run in Prisma
    project_id="proj-xyz789",    # routes spans to this project in Prisma
)

# Any OpenTelemetry span emitted from here on flows to Prisma.
from opentelemetry import trace

tracer = trace.get_tracer(__name__)
with tracer.start_as_current_span("review.case") as span:
    span.set_attribute("case.id", "CMP-2025-0001")
    ...

With PALQEE_PRISMA_API_KEY and PALQEE_PRISMA_OTEL_ENDPOINT set, the call collapses to configure_tracing(service_name="...").

What gets exported

Once configured, every span emitted by your process — from your own code or from any OpenTelemetry instrumentation library — flows to Prisma over OTLP HTTP. That includes spans from popular LLM instrumentation:

Provider / framework	Instrumentation
OpenAI, Anthropic Claude, Meta Llama, Bedrock, Gemini	OpenInference or OpenLLMetry
LangChain, LangGraph	OpenInference / OpenLLMetry
LlamaIndex	OpenInference
CrewAI, AutoGen, Haystack	OpenLLMetry
Custom code	opentelemetry-api spans

palqee-prisma-otel does not install instrumentation libraries — pick the ones you want and install them alongside.

Configuration

Parameter	Environment variable	Default	Description
api_key	PALQEE_PRISMA_API_KEY	required	Sent as x-api-key header to the Prisma Receiver
endpoint	PALQEE_PRISMA_OTEL_ENDPOINT	https://otel.prisma.ai:4318	OTLP HTTP endpoint of your Prisma deployment
service_name	—	"prisma-app"	OTEL service.name resource attribute
run_id	—	None	Attached as prisma.run.id for trace-to-run routing
project_id	—	None	Attached as prisma.project.id for trace-to-project routing

Full configuration reference: docs.palqee.com/docs/overview.

Features

• Authenticated OTLP export — x-api-key header pre-wired to the Prisma Receiver.
• HTTP transport — uses the standard opentelemetry-exporter-otlp-proto-http exporter.
• Batch span processing — efficient buffered export, no per-call network overhead.
• Resource attributes — prisma.run.id and prisma.project.id for server-side routing.
• Environment-first — runs in 12-factor / containerized deployments without code changes.

When to use this vs palqee-prisma-ai

Need	Reach for
Instrument an app, run validators, ship to Prisma in one call	palqee-prisma-ai
Send OpenTelemetry spans to Prisma from an existing OTEL setup	palqee-prisma-otel (this package)
Both	The SDK uses this wrapper under the hood — installing the SDK is enough

Links

• Documentation: docs.palqee.com/docs/overview
• High-level SDK: palqee-prisma-ai
• Platform client: palqee-prisma-client
• Website: palqee.com
• Support: [e-mail]

License

Apache 2.0 © Palqee Ltd.