Praison AI

observability

Observability diagnostics and trace verification CLI

praisonai obs

Observability diagnostics and management commands.

Usage

bash
# Via praisonai wrapper
praisonai obs [COMMAND] [OPTIONS]

# Standalone (no wrapper needed)
python -m praisonai_tools.observability.cli [COMMAND] [OPTIONS]

Commands

doctor

Run observability health checks.

bash
praisonai obs doctor
praisonai obs doctor --json
OptionDescription
--jsonOutput as JSON

Output:

CheckDescription
EnabledWhether observability is initialized
Active ProviderCurrently configured provider
ConnectionProvider connectivity status
Available ProvidersProviders with installed dependencies
Registered ProvidersAll known provider plugins

verify

Verify traces are recorded in the observability backend using the provider’s SDK.

bash
praisonai obs verify --provider langsmith --project "My First App"
praisonai obs verify --provider langsmith --project "My First App" --json
OptionDefaultDescription
--providerlangsmithProvider to verify
--projectdefaultProject name to check
--limit5Number of recent runs to check
--jsonOutput as JSON

Required environment variables:

bash
export LANGSMITH_API_KEY=lsv2_xxx
export LANGSMITH_ENDPOINT=https://api.smith.langchain.com  # or eu endpoint

Output:

The verify command checks each run for PraisonAI branding metadata (praisonai.version, praisonai.framework) and shows inputs/outputs status.

Examples

Quick Health Check

bash
python -m praisonai_tools.observability.cli doctor --json
json
{
  "enabled": true,
  "provider": "langsmith",
  "connection_status": true,
  "connection_message": "LangSmith API key configured"
}

Verify LangSmith Traces

bash
export LANGSMITH_API_KEY=lsv2_xxx
python -m praisonai_tools.observability.cli verify --project "My First App"

Programmatic Usage

python
from praisonai_tools.observability import obs

obs.init(provider="langsmith")
results = obs.doctor()
print(results)

What Gets Checked

  • Provider initialization status
  • API key configuration
  • Backend connectivity
  • Trace branding (praisonai.version, praisonai.framework)
  • Input/output data capture
  • Agent and workflow span recording

Related