[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M74Z7WF1Y8Q3VFAQPPPRAFMV":3,"skill-same:01M74Z7WF1Y8Q3VFAQPPPRAFMV":64},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":30,"commitSha":31,"skillDirectory":32,"forksCount":33,"repositoryHtmlUrl":34,"repositoryHomepageUrl":35,"ownerHtmlUrl":36,"ownerBlogUrl":37,"ownerDescription":38,"updatedAt":39,"install":40,"resources":53},"01M74Z7WF1Y8Q3VFAQPPPRAFMV","phoenix-observability",null,"Open-source AI observability platform for LLM tracing, evaluation, and monitoring. Use when debugging LLM applications with detailed traces, running evaluations on datasets, or monitoring production AI systems with real-time insights.","KalarisLabs\u002Fresearch-agent-skills","KalarisLabs","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F300878568?v=4","01M74Y7MT23GXXC56VM6TVSXZY","Kalarislabs","01M74Y7MTQM9REXZX4SNXWX95C",132,"MIT","https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fphoenix-observability\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: phoenix-observability\ndescription: Open-source AI observability platform for LLM tracing, evaluation, and monitoring. Use when debugging LLM applications with detailed traces, running evaluations on datasets, or monitoring production AI systems with real-time insights.\nlicense: MIT\nmetadata:\n  version: 1.0.0\n  category: ml-inference-and-ops\n  maintainer: Kalaris Labs\n  tags: Observability, Phoenix, Arize, Tracing, Evaluation, Monitoring, LLM Ops, OpenTelemetry\n  dependencies: arize-phoenix>=12.0.0\n---\n\n# Phoenix - AI Observability Platform\n\nOpen-source AI observability and evaluation platform for LLM applications with tracing, evaluation, datasets, experiments, and real-time monitoring.\n\n## When to use Phoenix\n\n**Use Phoenix when:**\n- Debugging LLM application issues with detailed traces\n- Running systematic evaluations on datasets\n- Monitoring production LLM systems in real-time\n- Building experiment pipelines for prompt\u002Fmodel comparison\n- Self-hosted observability without vendor lock-in\n\n**Key features:**\n- **Tracing**: OpenTelemetry-based trace collection for any LLM framework\n- **Evaluation**: LLM-as-judge evaluators for quality assessment\n- **Datasets**: Versioned test sets for regression testing\n- **Experiments**: Compare prompts, models, and configurations\n- **Playground**: Interactive prompt testing with multiple models\n- **Open-source**: Self-hosted with PostgreSQL or SQLite\n\n**Use alternatives instead:**\n- **LangSmith**: Managed platform with LangChain-first integration\n- **Weights & Biases**: Deep learning experiment tracking focus\n- **Arize Cloud**: Managed Phoenix with enterprise features\n- **MLflow**: General ML lifecycle, model registry focus\n\n## Quick start\n\n### Installation\n\n```bash\npip install arize-phoenix\n\n# With specific backends\npip install arize-phoenix[embeddings]  # Embedding analysis\npip install arize-phoenix-otel         # OpenTelemetry config\npip install arize-phoenix-evals        # Evaluation framework\npip install arize-phoenix-client       # Lightweight REST client\n```\n\n### Launch Phoenix server\n\n```python\nimport phoenix as px\n\n# Launch in notebook (ThreadServer mode)\nsession = px.launch_app()\n\n# View UI\nsession.view()  # Embedded iframe\nprint(session.url)  # http:\u002F\u002Flocalhost:6006\n```\n\n### Command-line server (production)\n\n```bash\n# Start Phoenix server\nphoenix serve\n\n# With PostgreSQL\nexport PHOENIX_SQL_DATABASE_URL=\"postgresql:\u002F\u002Fuser:pass@host\u002Fdb\"\nphoenix serve --port 6006\n```\n\n### Basic tracing\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.openai import OpenAIInstrumentor\n\n# Configure OpenTelemetry with Phoenix\ntracer_provider = register(\n    project_name=\"my-llm-app\",\n    endpoint=\"http:\u002F\u002Flocalhost:6006\u002Fv1\u002Ftraces\"\n)\n\n# Instrument OpenAI SDK\nOpenAIInstrumentor().instrument(tracer_provider=tracer_provider)\n\n# All OpenAI calls are now traced\nfrom openai import OpenAI\nclient = OpenAI()\nresponse = client.chat.completions.create(\n    model=\"gpt-4o\",\n    messages=[{\"role\": \"user\", \"content\": \"Hello!\"}]\n)\n```\n\n## Core concepts\n\n### Traces and spans\n\nA **trace** represents a complete execution flow, while **spans** are individual operations within that trace.\n\n```python\nfrom phoenix.otel import register\nfrom opentelemetry import trace\n\n# Setup tracing\ntracer_provider = register(project_name=\"my-app\")\ntracer = trace.get_tracer(__name__)\n\n# Create custom spans\nwith tracer.start_as_current_span(\"process_query\") as span:\n    span.set_attribute(\"input.value\", query)\n\n    # Child spans are automatically nested\n    with tracer.start_as_current_span(\"retrieve_context\"):\n        context = retriever.search(query)\n\n    with tracer.start_as_current_span(\"generate_response\"):\n        response = llm.generate(query, context)\n\n    span.set_attribute(\"output.value\", response)\n```\n\n### Projects\n\nProjects organize related traces:\n\n```python\nimport os\nos.environ[\"PHOENIX_PROJECT_NAME\"] = \"production-chatbot\"\n\n# Or per-trace\nfrom phoenix.otel import register\ntracer_provider = register(project_name=\"experiment-v2\")\n```\n\n## Framework instrumentation\n\n### OpenAI\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.openai import OpenAIInstrumentor\n\ntracer_provider = register()\nOpenAIInstrumentor().instrument(tracer_provider=tracer_provider)\n```\n\n### LangChain\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.langchain import LangChainInstrumentor\n\ntracer_provider = register()\nLangChainInstrumentor().instrument(tracer_provider=tracer_provider)\n\n# All LangChain operations traced\nfrom langchain_openai import ChatOpenAI\nllm = ChatOpenAI(model=\"gpt-4o\")\nresponse = llm.invoke(\"Hello!\")\n```\n\n### LlamaIndex\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.llama_index import LlamaIndexInstrumentor\n\ntracer_provider = register()\nLlamaIndexInstrumentor().instrument(tracer_provider=tracer_provider)\n```\n\n### Anthropic\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.anthropic import AnthropicInstrumentor\n\ntracer_provider = register()\nAnthropicInstrumentor().instrument(tracer_provider=tracer_provider)\n```\n\n## Evaluation framework\n\n### Built-in evaluators\n\n```python\nfrom phoenix.evals import (\n    OpenAIModel,\n    HallucinationEvaluator,\n    RelevanceEvaluator,\n    ToxicityEvaluator,\n    llm_classify\n)\n\n# Setup model for evaluation\neval_model = OpenAIModel(model=\"gpt-4o\")\n\n# Evaluate hallucination\nhallucination_eval = HallucinationEvaluator(eval_model)\nresults = hallucination_eval.evaluate(\n    input=\"What is the capital of France?\",\n    output=\"The capital of France is Paris.\",\n    reference=\"Paris is the capital of France.\"\n)\n```\n\n### Custom evaluators\n\n```python\nfrom phoenix.evals import llm_classify\n\n# Define custom evaluation\ndef evaluate_helpfulness(input_text, output_text):\n    template = \"\"\"\n    Evaluate if the response is helpful for the given question.\n\n    Question: {input}\n    Response: {output}\n\n    Is this response helpful? Answer 'helpful' or 'not_helpful'.\n    \"\"\"\n\n    result = llm_classify(\n        model=eval_model,\n        template=template,\n        input=input_text,\n        output=output_text,\n        rails=[\"helpful\", \"not_helpful\"]\n    )\n    return result\n```\n\n### Run evaluations on dataset\n\n```python\nfrom phoenix import Client\nfrom phoenix.evals import run_evals\n\nclient = Client()\n\n# Get spans to evaluate\nspans_df = client.get_spans_dataframe(\n    project_name=\"my-app\",\n    filter_condition=\"span_kind == 'LLM'\"\n)\n\n# Run evaluations\neval_results = run_evals(\n    dataframe=spans_df,\n    evaluators=[\n        HallucinationEvaluator(eval_model),\n        RelevanceEvaluator(eval_model)\n    ],\n    provide_explanation=True\n)\n\n# Log results back to Phoenix\nclient.log_evaluations(eval_results)\n```\n\n## Datasets and experiments\n\n### Create dataset\n\n```python\nfrom phoenix import Client\n\nclient = Client()\n\n# Create dataset\ndataset = client.create_dataset(\n    name=\"qa-test-set\",\n    description=\"QA evaluation dataset\"\n)\n\n# Add examples\nclient.add_examples_to_dataset(\n    dataset_name=\"qa-test-set\",\n    examples=[\n        {\n            \"input\": {\"question\": \"What is Python?\"},\n            \"output\": {\"answer\": \"A programming language\"}\n        },\n        {\n            \"input\": {\"question\": \"What is ML?\"},\n            \"output\": {\"answer\": \"Machine learning\"}\n        }\n    ]\n)\n```\n\n### Run experiment\n\n```python\nfrom phoenix import Client\nfrom phoenix.experiments import run_experiment\n\nclient = Client()\n\ndef my_model(input_data):\n    \"\"\"Your model function.\"\"\"\n    question = input_data[\"question\"]\n    return {\"answer\": generate_answer(question)}\n\ndef accuracy_evaluator(input_data, output, expected):\n    \"\"\"Custom evaluator.\"\"\"\n    return {\n        \"score\": 1.0 if expected[\"answer\"].lower() in output[\"answer\"].lower() else 0.0,\n        \"label\": \"correct\" if expected[\"answer\"].lower() in output[\"answer\"].lower() else \"incorrect\"\n    }\n\n# Run experiment\nresults = run_experiment(\n    dataset_name=\"qa-test-set\",\n    task=my_model,\n    evaluators=[accuracy_evaluator],\n    experiment_name=\"baseline-v1\"\n)\n\nprint(f\"Average accuracy: {results.aggregate_metrics['accuracy']}\")\n```\n\n## Client API\n\n### Query traces and spans\n\n```python\nfrom phoenix import Client\n\nclient = Client(endpoint=\"http:\u002F\u002Flocalhost:6006\")\n\n# Get spans as DataFrame\nspans_df = client.get_spans_dataframe(\n    project_name=\"my-app\",\n    filter_condition=\"span_kind == 'LLM'\",\n    limit=1000\n)\n\n# Get specific span\nspan = client.get_span(span_id=\"abc123\")\n\n# Get trace\ntrace = client.get_trace(trace_id=\"xyz789\")\n```\n\n### Log feedback\n\n```python\nfrom phoenix import Client\n\nclient = Client()\n\n# Log user feedback\nclient.log_annotation(\n    span_id=\"abc123\",\n    name=\"user_rating\",\n    annotator_kind=\"HUMAN\",\n    score=0.8,\n    label=\"helpful\",\n    metadata={\"comment\": \"Good response\"}\n)\n```\n\n### Export data\n\n```python\n# Export to pandas\ndf = client.get_spans_dataframe(project_name=\"my-app\")\n\n# Export traces\ntraces = client.list_traces(project_name=\"my-app\")\n```\n\n## Production deployment\n\n### Docker\n\n```bash\ndocker run -p 6006:6006 arizephoenix\u002Fphoenix:latest\n```\n\n### With PostgreSQL\n\n```bash\n# Set database URL\nexport PHOENIX_SQL_DATABASE_URL=\"postgresql:\u002F\u002Fuser:pass@host:5432\u002Fphoenix\"\n\n# Start server\nphoenix serve --host 0.0.0.0 --port 6006\n```\n\n### Environment variables\n\n| Variable | Description | Default |\n|----------|-------------|---------|\n| `PHOENIX_PORT` | HTTP server port | `6006` |\n| `PHOENIX_HOST` | Server bind address | `127.0.0.1` |\n| `PHOENIX_GRPC_PORT` | gRPC\u002FOTLP port | `4317` |\n| `PHOENIX_SQL_DATABASE_URL` | Database connection | SQLite temp |\n| `PHOENIX_WORKING_DIR` | Data storage directory | OS temp |\n| `PHOENIX_ENABLE_AUTH` | Enable authentication | `false` |\n| `PHOENIX_SECRET` | JWT signing secret | Required if auth enabled |\n\n### With authentication\n\n```bash\nexport PHOENIX_ENABLE_AUTH=true\nexport PHOENIX_SECRET=\"your-secret-key-min-32-chars\"\nexport PHOENIX_ADMIN_SECRET=\"admin-bootstrap-token\"\n\nphoenix serve\n```\n\n## Best practices\n\n1. **Use projects**: Separate traces by environment (dev\u002Fstaging\u002Fprod)\n2. **Add metadata**: Include user IDs, session IDs for debugging\n3. **Evaluate regularly**: Run automated evaluations in CI\u002FCD\n4. **Version datasets**: Track test set changes over time\n5. **Monitor costs**: Track token usage via Phoenix dashboards\n6. **Self-host**: Use PostgreSQL for production deployments\n\n## Common issues\n\n**Traces not appearing:**\n```python\nfrom phoenix.otel import register\n\n# Verify endpoint\ntracer_provider = register(\n    project_name=\"my-app\",\n    endpoint=\"http:\u002F\u002Flocalhost:6006\u002Fv1\u002Ftraces\"  # Correct endpoint\n)\n\n# Force flush\nfrom opentelemetry import trace\ntrace.get_tracer_provider().force_flush()\n```\n\n**High memory in notebook:**\n```python\n# Close session when done\nsession = px.launch_app()\n# ... do work ...\nsession.close()\npx.close_app()\n```\n\n**Database connection issues:**\n```bash\n# Verify PostgreSQL connection\npsql $PHOENIX_SQL_DATABASE_URL -c \"SELECT 1\"\n\n# Check Phoenix logs\nphoenix serve --log-level debug\n```\n\n## References\n\n- **[Advanced Usage](references\u002Fadvanced-usage.md)** - Custom evaluators, experiments, production setup\n- **[Troubleshooting](references\u002Ftroubleshooting.md)** - Common issues, debugging, performance\n\n## Resources\n\n- **Documentation**: https:\u002F\u002Fdocs.arize.com\u002Fphoenix\n- **Repository**: https:\u002F\u002Fgithub.com\u002FArize-ai\u002Fphoenix\n- **Docker Hub**: https:\u002F\u002Fhub.docker.com\u002Fr\u002Farizephoenix\u002Fphoenix\n- **Version**: 12.0.0+\n- **License**: Apache 2.0\n\n## Related skills\n\n- `langsmith-observability`: LLM observability platform for tracing, evaluation, and monitoring.\n- `chroma`: Open-source embedding database for AI applications.\n- `speculative-decoding`: Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques.\n","# Phoenix - AI Observability Platform\n\nOpen-source AI observability and evaluation platform for LLM applications with tracing, evaluation, datasets, experiments, and real-time monitoring.\n\n## When to use Phoenix\n\n**Use Phoenix when:**\n- Debugging LLM application issues with detailed traces\n- Running systematic evaluations on datasets\n- Monitoring production LLM systems in real-time\n- Building experiment pipelines for prompt\u002Fmodel comparison\n- Self-hosted observability without vendor lock-in\n\n**Key features:**\n- **Tracing**: OpenTelemetry-based trace collection for any LLM framework\n- **Evaluation**: LLM-as-judge evaluators for quality assessment\n- **Datasets**: Versioned test sets for regression testing\n- **Experiments**: Compare prompts, models, and configurations\n- **Playground**: Interactive prompt testing with multiple models\n- **Open-source**: Self-hosted with PostgreSQL or SQLite\n\n**Use alternatives instead:**\n- **LangSmith**: Managed platform with LangChain-first integration\n- **Weights & Biases**: Deep learning experiment tracking focus\n- **Arize Cloud**: Managed Phoenix with enterprise features\n- **MLflow**: General ML lifecycle, model registry focus\n\n## Quick start\n\n### Installation\n\n```bash\npip install arize-phoenix\n\n# With specific backends\npip install arize-phoenix[embeddings]  # Embedding analysis\npip install arize-phoenix-otel         # OpenTelemetry config\npip install arize-phoenix-evals        # Evaluation framework\npip install arize-phoenix-client       # Lightweight REST client\n```\n\n### Launch Phoenix server\n\n```python\nimport phoenix as px\n\n# Launch in notebook (ThreadServer mode)\nsession = px.launch_app()\n\n# View UI\nsession.view()  # Embedded iframe\nprint(session.url)  # http:\u002F\u002Flocalhost:6006\n```\n\n### Command-line server (production)\n\n```bash\n# Start Phoenix server\nphoenix serve\n\n# With PostgreSQL\nexport PHOENIX_SQL_DATABASE_URL=\"postgresql:\u002F\u002Fuser:pass@host\u002Fdb\"\nphoenix serve --port 6006\n```\n\n### Basic tracing\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.openai import OpenAIInstrumentor\n\n# Configure OpenTelemetry with Phoenix\ntracer_provider = register(\n    project_name=\"my-llm-app\",\n    endpoint=\"http:\u002F\u002Flocalhost:6006\u002Fv1\u002Ftraces\"\n)\n\n# Instrument OpenAI SDK\nOpenAIInstrumentor().instrument(tracer_provider=tracer_provider)\n\n# All OpenAI calls are now traced\nfrom openai import OpenAI\nclient = OpenAI()\nresponse = client.chat.completions.create(\n    model=\"gpt-4o\",\n    messages=[{\"role\": \"user\", \"content\": \"Hello!\"}]\n)\n```\n\n## Core concepts\n\n### Traces and spans\n\nA **trace** represents a complete execution flow, while **spans** are individual operations within that trace.\n\n```python\nfrom phoenix.otel import register\nfrom opentelemetry import trace\n\n# Setup tracing\ntracer_provider = register(project_name=\"my-app\")\ntracer = trace.get_tracer(__name__)\n\n# Create custom spans\nwith tracer.start_as_current_span(\"process_query\") as span:\n    span.set_attribute(\"input.value\", query)\n\n    # Child spans are automatically nested\n    with tracer.start_as_current_span(\"retrieve_context\"):\n        context = retriever.search(query)\n\n    with tracer.start_as_current_span(\"generate_response\"):\n        response = llm.generate(query, context)\n\n    span.set_attribute(\"output.value\", response)\n```\n\n### Projects\n\nProjects organize related traces:\n\n```python\nimport os\nos.environ[\"PHOENIX_PROJECT_NAME\"] = \"production-chatbot\"\n\n# Or per-trace\nfrom phoenix.otel import register\ntracer_provider = register(project_name=\"experiment-v2\")\n```\n\n## Framework instrumentation\n\n### OpenAI\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.openai import OpenAIInstrumentor\n\ntracer_provider = register()\nOpenAIInstrumentor().instrument(tracer_provider=tracer_provider)\n```\n\n### LangChain\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.langchain import LangChainInstrumentor\n\ntracer_provider = register()\nLangChainInstrumentor().instrument(tracer_provider=tracer_provider)\n\n# All LangChain operations traced\nfrom langchain_openai import ChatOpenAI\nllm = ChatOpenAI(model=\"gpt-4o\")\nresponse = llm.invoke(\"Hello!\")\n```\n\n### LlamaIndex\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.llama_index import LlamaIndexInstrumentor\n\ntracer_provider = register()\nLlamaIndexInstrumentor().instrument(tracer_provider=tracer_provider)\n```\n\n### Anthropic\n\n```python\nfrom phoenix.otel import register\nfrom openinference.instrumentation.anthropic import AnthropicInstrumentor\n\ntracer_provider = register()\nAnthropicInstrumentor().instrument(tracer_provider=tracer_provider)\n```\n\n## Evaluation framework\n\n### Built-in evaluators\n\n```python\nfrom phoenix.evals import (\n    OpenAIModel,\n    HallucinationEvaluator,\n    RelevanceEvaluator,\n    ToxicityEvaluator,\n    llm_classify\n)\n\n# Setup model for evaluation\neval_model = OpenAIModel(model=\"gpt-4o\")\n\n# Evaluate hallucination\nhallucination_eval = HallucinationEvaluator(eval_model)\nresults = hallucination_eval.evaluate(\n    input=\"What is the capital of France?\",\n    output=\"The capital of France is Paris.\",\n    reference=\"Paris is the capital of France.\"\n)\n```\n\n### Custom evaluators\n\n```python\nfrom phoenix.evals import llm_classify\n\n# Define custom evaluation\ndef evaluate_helpfulness(input_text, output_text):\n    template = \"\"\"\n    Evaluate if the response is helpful for the given question.\n\n    Question: {input}\n    Response: {output}\n\n    Is this response helpful? Answer 'helpful' or 'not_helpful'.\n    \"\"\"\n\n    result = llm_classify(\n        model=eval_model,\n        template=template,\n        input=input_text,\n        output=output_text,\n        rails=[\"helpful\", \"not_helpful\"]\n    )\n    return result\n```\n\n### Run evaluations on dataset\n\n```python\nfrom phoenix import Client\nfrom phoenix.evals import run_evals\n\nclient = Client()\n\n# Get spans to evaluate\nspans_df = client.get_spans_dataframe(\n    project_name=\"my-app\",\n    filter_condition=\"span_kind == 'LLM'\"\n)\n\n# Run evaluations\neval_results = run_evals(\n    dataframe=spans_df,\n    evaluators=[\n        HallucinationEvaluator(eval_model),\n        RelevanceEvaluator(eval_model)\n    ],\n    provide_explanation=True\n)\n\n# Log results back to Phoenix\nclient.log_evaluations(eval_results)\n```\n\n## Datasets and experiments\n\n### Create dataset\n\n```python\nfrom phoenix import Client\n\nclient = Client()\n\n# Create dataset\ndataset = client.create_dataset(\n    name=\"qa-test-set\",\n    description=\"QA evaluation dataset\"\n)\n\n# Add examples\nclient.add_examples_to_dataset(\n    dataset_name=\"qa-test-set\",\n    examples=[\n        {\n            \"input\": {\"question\": \"What is Python?\"},\n            \"output\": {\"answer\": \"A programming language\"}\n        },\n        {\n            \"input\": {\"question\": \"What is ML?\"},\n            \"output\": {\"answer\": \"Machine learning\"}\n        }\n    ]\n)\n```\n\n### Run experiment\n\n```python\nfrom phoenix import Client\nfrom phoenix.experiments import run_experiment\n\nclient = Client()\n\ndef my_model(input_data):\n    \"\"\"Your model function.\"\"\"\n    question = input_data[\"question\"]\n    return {\"answer\": generate_answer(question)}\n\ndef accuracy_evaluator(input_data, output, expected):\n    \"\"\"Custom evaluator.\"\"\"\n    return {\n        \"score\": 1.0 if expected[\"answer\"].lower() in output[\"answer\"].lower() else 0.0,\n        \"label\": \"correct\" if expected[\"answer\"].lower() in output[\"answer\"].lower() else \"incorrect\"\n    }\n\n# Run experiment\nresults = run_experiment(\n    dataset_name=\"qa-test-set\",\n    task=my_model,\n    evaluators=[accuracy_evaluator],\n    experiment_name=\"baseline-v1\"\n)\n\nprint(f\"Average accuracy: {results.aggregate_metrics['accuracy']}\")\n```\n\n## Client API\n\n### Query traces and spans\n\n```python\nfrom phoenix import Client\n\nclient = Client(endpoint=\"http:\u002F\u002Flocalhost:6006\")\n\n# Get spans as DataFrame\nspans_df = client.get_spans_dataframe(\n    project_name=\"my-app\",\n    filter_condition=\"span_kind == 'LLM'\",\n    limit=1000\n)\n\n# Get specific span\nspan = client.get_span(span_id=\"abc123\")\n\n# Get trace\ntrace = client.get_trace(trace_id=\"xyz789\")\n```\n\n### Log feedback\n\n```python\nfrom phoenix import Client\n\nclient = Client()\n\n# Log user feedback\nclient.log_annotation(\n    span_id=\"abc123\",\n    name=\"user_rating\",\n    annotator_kind=\"HUMAN\",\n    score=0.8,\n    label=\"helpful\",\n    metadata={\"comment\": \"Good response\"}\n)\n```\n\n### Export data\n\n```python\n# Export to pandas\ndf = client.get_spans_dataframe(project_name=\"my-app\")\n\n# Export traces\ntraces = client.list_traces(project_name=\"my-app\")\n```\n\n## Production deployment\n\n### Docker\n\n```bash\ndocker run -p 6006:6006 arizephoenix\u002Fphoenix:latest\n```\n\n### With PostgreSQL\n\n```bash\n# Set database URL\nexport PHOENIX_SQL_DATABASE_URL=\"postgresql:\u002F\u002Fuser:pass@host:5432\u002Fphoenix\"\n\n# Start server\nphoenix serve --host 0.0.0.0 --port 6006\n```\n\n### Environment variables\n\n| Variable | Description | Default |\n|----------|-------------|---------|\n| `PHOENIX_PORT` | HTTP server port | `6006` |\n| `PHOENIX_HOST` | Server bind address | `127.0.0.1` |\n| `PHOENIX_GRPC_PORT` | gRPC\u002FOTLP port | `4317` |\n| `PHOENIX_SQL_DATABASE_URL` | Database connection | SQLite temp |\n| `PHOENIX_WORKING_DIR` | Data storage directory | OS temp |\n| `PHOENIX_ENABLE_AUTH` | Enable authentication | `false` |\n| `PHOENIX_SECRET` | JWT signing secret | Required if auth enabled |\n\n### With authentication\n\n```bash\nexport PHOENIX_ENABLE_AUTH=true\nexport PHOENIX_SECRET=\"your-secret-key-min-32-chars\"\nexport PHOENIX_ADMIN_SECRET=\"admin-bootstrap-token\"\n\nphoenix serve\n```\n\n## Best practices\n\n1. **Use projects**: Separate traces by environment (dev\u002Fstaging\u002Fprod)\n2. **Add metadata**: Include user IDs, session IDs for debugging\n3. **Evaluate regularly**: Run automated evaluations in CI\u002FCD\n4. **Version datasets**: Track test set changes over time\n5. **Monitor costs**: Track token usage via Phoenix dashboards\n6. **Self-host**: Use PostgreSQL for production deployments\n\n## Common issues\n\n**Traces not appearing:**\n```python\nfrom phoenix.otel import register\n\n# Verify endpoint\ntracer_provider = register(\n    project_name=\"my-app\",\n    endpoint=\"http:\u002F\u002Flocalhost:6006\u002Fv1\u002Ftraces\"  # Correct endpoint\n)\n\n# Force flush\nfrom opentelemetry import trace\ntrace.get_tracer_provider().force_flush()\n```\n\n**High memory in notebook:**\n```python\n# Close session when done\nsession = px.launch_app()\n# ... do work ...\nsession.close()\npx.close_app()\n```\n\n**Database connection issues:**\n```bash\n# Verify PostgreSQL connection\npsql $PHOENIX_SQL_DATABASE_URL -c \"SELECT 1\"\n\n# Check Phoenix logs\nphoenix serve --log-level debug\n```\n\n## References\n\n- **[Advanced Usage](references\u002Fadvanced-usage.md)** - Custom evaluators, experiments, production setup\n- **[Troubleshooting](references\u002Ftroubleshooting.md)** - Common issues, debugging, performance\n\n## Resources\n\n- **Documentation**: https:\u002F\u002Fdocs.arize.com\u002Fphoenix\n- **Repository**: https:\u002F\u002Fgithub.com\u002FArize-ai\u002Fphoenix\n- **Docker Hub**: https:\u002F\u002Fhub.docker.com\u002Fr\u002Farizephoenix\u002Fphoenix\n- **Version**: 12.0.0+\n- **License**: Apache 2.0\n\n## Related skills\n\n- `langsmith-observability`: LLM observability platform for tracing, evaluation, and monitoring.\n- `chroma`: Open-source embedding database for AI applications.\n- `speculative-decoding`: Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques.\n",{"name":5,"license":15,"metadata":24,"description":7},{"tags":25,"version":26,"category":27,"maintainer":28,"dependencies":29},"Observability, Phoenix, Arize, Tracing, Evaluation, Monitoring, LLM Ops, OpenTelemetry","1.0.0","ml-inference-and-ops","Kalaris Labs","arize-phoenix>=12.0.0","main","29fe36d26127205057d5e0ed8425f6b0e9a8c13a","skills\u002Fphoenix-observability",16,"https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills","https:\u002F\u002Fdocs.kalarislabs.com\u002Fresearch-agent-skills","https:\u002F\u002Fgithub.com\u002FKalarisLabs","kalarislabs.com","Applied AI lab building the infrastructure for the agentic era","2026-11-14 17:49:08",{"source":8,"skillName":5,"commands":41,"prompt":51,"zipAvailable":52},[42,45,48],{"tool":43,"command":44},"npx","npx skills add KalarisLabs\u002Fresearch-agent-skills --skill phoenix-observability",{"tool":46,"command":47},"bunx","bunx skills add KalarisLabs\u002Fresearch-agent-skills --skill phoenix-observability",{"tool":49,"command":50},"pnpm","pnpm dlx skills add KalarisLabs\u002Fresearch-agent-skills --skill phoenix-observability","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fphoenix-observability\u002FSKILL.md",true,[54,60],{"path":55,"resourceType":56,"resourceTypeDesc":57,"sizeBytes":58,"sourceHtmlUrl":59},"references\u002Fadvanced-usage.md","REFERENCE","参考文档",15277,"https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fphoenix-observability\u002Freferences\u002Fadvanced-usage.md",{"path":61,"resourceType":56,"resourceTypeDesc":57,"sizeBytes":62,"sourceHtmlUrl":63},"references\u002Ftroubleshooting.md",10944,"https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fphoenix-observability\u002Freferences\u002Ftroubleshooting.md",[65,72,79,86,93,100],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M74Z8MGT305R11SNY5G28QH5","plos","Prepare manuscripts for PLOS journals (PLOS ONE, PLOS Biology, PLOS Computational Biology, PLOS Genetics, PLOS Medicine, PLOS Pathogens, PLOS Neglected Tropical Diseases, PLOS Global Public Health and others), covering structure, abstract and author summary, Vancouver-style references, the mandatory data availability policy, ethics and financial disclosure statements, figure requirements (PACE), the PLOS LaTeX template, reporting guidelines and PLOS ONE's publication criteria. Use when targeting any PLOS journal or checking compliance with PLOS policies.","https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fplos\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":19},"01M74Z8CCN9APXYHNQ95V1EBGT","pkpd-modeling","Pharmacokinetic and pharmacodynamic modelling and simulation - non-compartmental analysis, compartmental and population PK, PK\u002FPD and exposure-response, TMDD, PBPK orientation, bioequivalence, allometric scaling and first-in-human dose, drug interaction prediction, and Bayesian therapeutic drug monitoring. Use when analysing concentration-time data, deriving exposure metrics, fitting PK or PD models, or evaluating dosing regimens. Triggers include \"pharmacokinetics\", \"pharmacodynamics\", \"PK\u002FPD\", \"NCA\", \"non-compartmental\", \"AUC\", \"Cmax\", \"lambda z\", \"half-life\", \"clearance\", \"volume of distribution\", \"compartmental model\", \"population PK\", \"popPK\", \"NONMEM\", \"nlmixr2\", \"Pharmpy\", \"Monolix\", \"exposure-response\", \"Emax\", \"EC50\", \"indirect response\", \"effect compartment\", \"TMDD\", \"PBPK\", \"bioequivalence\", \"RSABE\", \"ABEL\", \"allometric scaling\", \"first-in-human\", \"MABEL\", \"drug-drug interaction\", \"DDI\", \"ICH M12\", \"concentration-QTc\", \"therapeutic drug monitoring\", \"MIPD\", and \"dosing regimen\".","https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpkpd-modeling\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":19},"01M74Z86YHJJ1FM3X87BHN76KZ","pinecone","Guides use of Pinecone, a managed serverless vector database, through its Python client and the LangChain and LlamaIndex integrations. Covers creating indexes, upserting and querying vectors, metadata filtering, namespaces, hybrid dense and sparse search, index management, and deleting vectors. Use when building a production RAG system on a hosted vector store, adding semantic search or recommendations without running infrastructure, isolating per-user or per-tenant data with namespaces, combining dense and sparse vectors in one query, or filtering results by metadata. Do not use for self-hosted or local stores (Chroma, Weaviate) or offline similarity search (FAISS).","https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpinecone\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":19},"01M74Z80T6E89SAJ8MYDD0767Q","pi-agent","Build with and use Pi, the minimal terminal coding harness. Use for installing Pi, configuring providers\u002Fmodels\u002Fsettings\u002Fenvironment variables, creating Pi skills\u002Fextensions\u002Fpackages\u002Fthemes\u002Fprompt templates, embedding Pi through the SDK, integrating over RPC or JSON event streams, parsing sessions, running local models through the llama.cpp router, developing custom Pi providers and TUI components, or using ecosystem packages such as pi-subagents (delegation\u002Forchestration), pi-mcp-adapter (MCP servers), pi-interview (interactive forms), and pi-web-access (web search, fetching, video understanding).","https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpi-agent\u002FSKILL.md",[],[],{"skillNo":94,"name":95,"nameZh":6,"description":96,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":19},"01M74Z7HNEH1A2AVYR3G1X2QV6","peft-fine-tuning","Fine-tunes LLMs with Hugging Face PEFT, using LoRA, QLoRA, IA3, AdaLoRA, prefix tuning, and prompt tuning so that under 1% of parameters are trained. Covers rank, alpha, and target module selection, loading and merging adapters, multi-adapter serving, and integration with TRL SFTTrainer, Axolotl, and vLLM. Use when fine-tuning 7B-70B models on limited GPU memory, when running QLoRA on a single 24GB GPU, when choosing LoRA rank and alpha, when merging or swapping adapters on one base model, or when debugging CUDA OOM or an adapter that does not apply. Not for full fine-tuning of small models under 1B parameters or cases needing all weights updated.","https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpeft-fine-tuning\u002FSKILL.md",[],[],{"skillNo":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":19},"01M74Z7CDPXF7EGG2MNSGRXZ4Z","peer-review","Prepare evidence-bounded, constructive peer-review drafts and structured manuscript assessments. Use for authorized review of scientific manuscripts, protocols, preprints, or research proposals; reporting-guideline selection; claim–evidence checks; methods, statistics, reproducibility, ethics, figure\u002Ftable, and citation critique; or revision-response planning.","https:\u002F\u002Fgithub.com\u002FKalarisLabs\u002Fresearch-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpeer-review\u002FSKILL.md",[],[]]