Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.praison.ai/llms.txt

Use this file to discover all available pages before exploring further.

Observability Hooks provide a centralized entry point for initializing observability providers (AgentOps, future Langfuse, W&B) in PraisonAI and custom framework adapters. praisonai.observability.hooks.init_observability(framework_tag, *, tags=None) is now a public hook, called automatically by the orchestrator and intended to be called by custom framework adapters’ setup() method.

​
Quick Start

1

Default behaviour (just set the env var)

export AGENTOPS_API_KEY=xxx

from praisonai.agents_generator import AgentsGenerator

# init_observability is called for you, with framework_tag = resolved adapter name
gen = AgentsGenerator("agents.yaml", "crewai", config_list=[...])
gen.generate_crew_and_kickoff()
2

Custom tags from a custom adapter

from praisonai.framework_adapters.base import BaseFrameworkAdapter
from praisonai.observability.hooks import init_observability

class MyAdapter(BaseFrameworkAdapter):
    name = "myframework"

    def setup(self, *, framework_tag: str) -> None:
        # Add extra tags for this run
        init_observability(framework_tag, tags=["tenant=acme", "experiment=foo"])
3

Branch on availability

from praisonai.observability.hooks import AGENTOPS_AVAILABLE

if AGENTOPS_AVAILABLE:
    ...  # do extra setup

​
How It Works

init_observability(framework_tag, *, tags=None) centralizes observability initialization:
  • Auto-call site: orchestrator calls init_observability(adapter.name) immediately after assert_framework_available(...) and before adapter.setup(...)
  • AgentOps init guard: agentops.init(...) only fires if both (a) agentops is importable, and (b) AGENTOPS_API_KEY is set in the env
  • Failure mode: ImportError (no agentops) is logged at DEBUG; any other exception is logged at WARNING and never propagated
  • AGENTOPS_AVAILABLE module-level boolean for code that wants to short-circuit instrumentation entirely
The hook also leaves room for future providers (the source already has placeholder comments for _init_langfuse and _init_wandb), so users may want to know the surface area.

​
Configuration

ParameterTypeDefaultDescription
framework_tagstrrequiredPrimary tag (e.g. "crewai", "autogen_v4"). Becomes the first entry in default_tags passed to agentops.init.
tagslist[str] | NoneNoneExtra tags appended after framework_tag.

​
Best Practices

The orchestrator calls init_observability(adapter.name) once per run. If you call it again from setup(), you’ll re-init with your tags (last call wins for AgentOps). Use this for run-scoped tags only:
def setup(self, *, framework_tag: str) -> None:
    # Good - adds run-specific context
    init_observability(framework_tag, tags=[
        f"tenant={self.tenant_id}",
        f"experiment={self.experiment_name}"
    ])
Don’t import agentops at the top of your adapter — gate it behind AGENTOPS_AVAILABLE or rely on the hook to no-op silently:
# ✅ Good - use the hook or check availability
from praisonai.observability.hooks import AGENTOPS_AVAILABLE, init_observability

if AGENTOPS_AVAILABLE:
    # Safe to do AgentOps-specific setup
    pass

# ❌ Bad - direct import can fail
import agentops  # May fail if not installed
New providers (Langfuse, W&B, etc.) will be added inside _init_<provider> helpers in praisonai/observability/hooks.py — calling init_observability(...) will automatically pick them up; you don’t need to update adapter code:
# Future providers will be added automatically
def init_observability(framework_tag, *, tags=None):
    _init_agentops(framework_tag, tags or [])
    # _init_langfuse(framework_tag, tags)    # Future
    # _init_wandb(framework_tag, tags)       # Future

AgentOps

AgentOps integration documentation

Framework Adapter Plugins

How to create custom framework adapters