Open Telemetry

[MervinPraison]

No description provided.

[MervinPraison]

@claude

why i am not seeing the telemetry posted to posthog ?
previously it was working and nwo its not working

❯ python telemetry_minimal.py
01:54:02 - LiteLLM:DEBUG: http_handler.py:530 - Using AiohttpTransport...
[01:54:02] DEBUG    [01:54:02] http_handler.py:530 DEBUG Using  http_handler.py:530
                    AiohttpTransport...                                            
01:54:02 - LiteLLM:DEBUG: http_handler.py:554 - Creating AiohttpTransport...
           DEBUG    [01:54:02] http_handler.py:554 DEBUG        http_handler.py:554
                    Creating AiohttpTransport...                                   
           DEBUG    [01:54:02] _trace.py:47 DEBUG connect_tcp.started  _trace.py:47
                    host='raw.githubusercontent.com' port=443                      
                    local_address=None timeout=5 socket_options=None               
           DEBUG    [01:54:02] _trace.py:47 DEBUG connect_tcp.complete _trace.py:47
                    return_value=<httpcore._backends.sync.SyncStream               
                    object at 0x1133e2510>                                         
           DEBUG    [01:54:02] _trace.py:47 DEBUG start_tls.started    _trace.py:47
                    ssl_context=<ssl.SSLContext object at 0x112e4f1d0>             
                    server_hostname='raw.githubusercontent.com'                    
                    timeout=5                                                      
           DEBUG    [01:54:02] _trace.py:47 DEBUG start_tls.complete   _trace.py:47
                    return_value=<httpcore._backends.sync.SyncStream               
                    object at 0x1137b11d0>                                         
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    send_request_headers.started request=<Request                  
                    [b'GET']>                                                      
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    send_request_headers.complete                                  
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    send_request_body.started request=<Request                     
                    [b'GET']>                                                      
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    send_request_body.complete                                     
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    receive_response_headers.started request=<Request              
                    [b'GET']>                                                      
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    receive_response_headers.complete                              
                    return_value=(b'HTTP/1.1', 200, b'OK',                         
                    [(b'Connection', b'keep-alive'),                               
                    (b'Content-Length', b'24480'), (b'Cache-Control',              
                    b'max-age=300'), (b'Content-Security-Policy',                  
                    b"default-src 'none'; style-src 'unsafe-inline';               
                    sandbox"), (b'Content-Type', b'text/plain;                     
                    charset=utf-8'), (b'ETag',                                     
                    b'W/"7d823a65ac55a83403e582c94a65fa7800f01d7f1389c             
                    8bc0e6cc6f4b9a552c0"'),                                        
                    (b'Strict-Transport-Security',                                 
                    b'max-age=31536000'), (b'X-Content-Type-Options',              
                    b'nosniff'), (b'X-Frame-Options', b'deny'),                    
                    (b'X-XSS-Protection', b'1; mode=block'),                       
                    (b'X-GitHub-Request-Id',                                       
                    b'D346:16501D:1DBBF:3E555:68432E67'),                          
                    (b'Content-Encoding', b'gzip'), (b'Accept-Ranges',             
                    b'bytes'), (b'Date', b'Sat, 07 Jun 2025 00:54:02               
                    GMT'), (b'Via', b'1.1 varnish'), (b'X-Served-By',              
                    b'cache-lcy-eglc8600076-LCY'), (b'X-Cache',                    
                    b'HIT'), (b'X-Cache-Hits', b'8'), (b'X-Timer',                 
                    b'S1749257642.478383,VS0,VE0'), (b'Vary',                      
                    b'Authorization,Accept-Encoding'),                             
                    (b'Access-Control-Allow-Origin', b'*'),                        
                    (b'Cross-Origin-Resource-Policy',                              
                    b'cross-origin'), (b'X-Fastly-Request-ID',                     
                    b'c190b23b3b8a70cc33ca6a5e7322bcf3d90bc6e4'),                  
                    (b'Expires', b'Sat, 07 Jun 2025 00:59:02 GMT'),                
                    (b'Source-Age', b'299')])                                      
           INFO     [01:54:02] _client.py:1025 INFO HTTP Request:   _client.py:1025
                    GET                                                            
                    https://raw.githubusercontent.com/BerriAI/litel                
                    lm/main/model_prices_and_context_window.json                   
                    "HTTP/1.1 200 OK"                                              
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    receive_response_body.started request=<Request                 
                    [b'GET']>                                                      
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    receive_response_body.complete                                 
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    response_closed.started                                        
           DEBUG    [01:54:02] _trace.py:47 DEBUG                      _trace.py:47
                    response_closed.complete                                       
           DEBUG    [01:54:02] _trace.py:47 DEBUG close.started        _trace.py:47
           DEBUG    [01:54:02] _trace.py:47 DEBUG close.complete       _trace.py:47
/Users/praison/praisonai-package/src/praisonai-agents/venv/lib/python3.13/site-packages/pydantic/_internal/_config.py:295: PydanticDeprecatedSince20: Support for class-based `config` is deprecated, use ConfigDict instead. Deprecated in Pydantic V2.0 to be removed in V3.0. See Pydantic V2 Migration Guide at https://errors.pydantic.dev/2.10/migration/
  warnings.warn(DEPRECATION_MESSAGE, DeprecationWarning)
01:54:02 - LiteLLM:DEBUG: http_handler.py:530 - Using AiohttpTransport...
           DEBUG    [01:54:02] http_handler.py:530 DEBUG Using  http_handler.py:530
                    AiohttpTransport...                                            
01:54:02 - LiteLLM:DEBUG: http_handler.py:554 - Creating AiohttpTransport...
           DEBUG    [01:54:02] http_handler.py:554 DEBUG        http_handler.py:554
                    Creating AiohttpTransport...                                   
01:54:02 - LiteLLM:DEBUG: litellm_logging.py:169 - [Non-Blocking] Unable to import GenericAPILogger - LiteLLM Enterprise Feature - No module named 'litellm_enterprise'
           DEBUG    [01:54:02] litellm_logging.py:169 DEBUG  litellm_logging.py:169
                    [Non-Blocking] Unable to import                                
                    GenericAPILogger - LiteLLM Enterprise                          
                    Feature - No module named                                      
                    'litellm_enterprise'                                           
01:54:02 - LiteLLM:DEBUG: transformation.py:17 - [Non-Blocking] Unable to import _ENTERPRISE_ResponsesSessionHandler - LiteLLM Enterprise Feature - No module named 'litellm_enterprise'
           DEBUG    [01:54:02] transformation.py:17 DEBUG      transformation.py:17
                    [Non-Blocking] Unable to import                                
                    _ENTERPRISE_ResponsesSessionHandler -                          
                    LiteLLM Enterprise Feature - No module                         
                    named 'litellm_enterprise'                                     
01:54:02 - LiteLLM:DEBUG: http_handler.py:530 - Using AiohttpTransport...
           DEBUG    [01:54:02] http_handler.py:530 DEBUG Using  http_handler.py:530
                    AiohttpTransport...                                            
01:54:02 - LiteLLM:DEBUG: http_handler.py:554 - Creating AiohttpTransport...
           DEBUG    [01:54:02] http_handler.py:554 DEBUG        http_handler.py:554
                    Creating AiohttpTransport...                                   
01:54:02 - LiteLLM:DEBUG: http_handler.py:530 - Using AiohttpTransport...
           DEBUG    [01:54:02] http_handler.py:530 DEBUG Using  http_handler.py:530
                    AiohttpTransport...                                            
01:54:02 - LiteLLM:DEBUG: http_handler.py:554 - Creating AiohttpTransport...
           DEBUG    [01:54:02] http_handler.py:554 DEBUG        http_handler.py:554
                    Creating AiohttpTransport...                                   
           DEBUG    [01:54:02] telemetry.py:81 DEBUG Telemetry      telemetry.py:81
                    enabled with session 7cb1e713629a7b0f                          
           DEBUG    [01:54:02] telemetry.py:178 DEBUG Feature      telemetry.py:178
                    usage tracked: workflow_sequential                             
Running workflow with telemetry...
           DEBUG    [01:54:02] process.py:21 DEBUG === Initializing   process.py:21
                    Process ===                                                    
           DEBUG    [01:54:02] process.py:22 DEBUG Number of tasks: 1 process.py:22
           DEBUG    [01:54:02] process.py:23 DEBUG Number of agents:  process.py:23
                    1                                                              
           DEBUG    [01:54:02] process.py:24 DEBUG Manager LLM:       process.py:24
                    gpt-4o                                                         
           DEBUG    [01:54:02] process.py:25 DEBUG Verbose mode: 0    process.py:25
           DEBUG    [01:54:02] process.py:26 DEBUG Max iterations: 10 process.py:26
           DEBUG    [01:54:02] agent.py:1050 DEBUG Agent.chat         agent.py:1050
                    parameters: {                                                  
                      "prompt": "\nYou need to do the following task:              
                    Calculate 2 + 2.\nExpected Output: The                         
                    sum.\nPlease provide only t...",                               
                      "temperature": 0.2,                                          
                      "tools": null,                                               
                      "output_json": null,                                         
                      "output_pydantic": null,                                     
                      "reasoning_steps": false,                                    
                      "agent_name": "Calculator",                                  
                      "agent_role": "Math Expert",                                 
                      "agent_goal": "Perform calculations"                         
                    }                                                              
╭─ Agent Info ────────────────────────────────────────────────────────────────────╮
│                                                                                 │
│  👤 Agent: Calculator                                                           │
│  Role: Math Expert                                                              │
│                                                                                 │
╰─────────────────────────────────────────────────────────────────────────────────╯
╭────────────────────────────────── Instruction ──────────────────────────────────╮
│ Agent Calculator is processing prompt:                                          │
│ You need to do the following task: Calculate 2 + 2.                             │
│ Expected Output: The sum.                                                       │
│ Please provide only the final result of your work. Do not add any conversation  │
│ or extra explanation.                                                           │
╰─────────────────────────────────────────────────────────────────────────────────╯
           DEBUG    [01:54:02] agent.py:879 DEBUG Calculator sending   agent.py:879
                    messages to LLM: [{'role': 'system', 'content':                
                    'You are a helpful math expert.\n\nYour Role: Math             
                    Expert\n\nYour Goal: Perform calculations\n                    
                    '}, {'role': 'user', 'content': '\nYou need to do              
                    the following task: Calculate 2 + 2.\nExpected                 
                    Output: The sum.\nPlease provide only the final                
                    result of your work. Do not add any conversation               
                    or extra explanation.'}]                                       
[01:54:04] DEBUG    [01:54:04] main.py:224 DEBUG Empty content in       main.py:224
                    display_generating, returning early                            

           DEBUG    [01:54:04] agent.py:1212 DEBUG Agent Calculator   agent.py:1212
                    final response: 4                                              
Response generated in 1.5s
╭───────────────────────────────────── Task ──────────────────────────────────────╮
│ You need to do the following task: Calculate 2 + 2. Expected Output: The sum.   │
│ Please provide only the final result of your work. Do not add any conversation  │
│ or extra explanation.                                                           │
╰─────────────────────────────────────────────────────────────────────────────────╯
╭─────────────────────────────────── Response ────────────────────────────────────╮
│ 4                                                                               │
╰─────────────────────────────────────────────────────────────────────────────────╯
           DEBUG    [01:54:04] telemetry.py:119 DEBUG Agent        telemetry.py:119
                    execution tracked: success=True                                
           DEBUG    [01:54:04] telemetry.py:134 DEBUG Task         telemetry.py:134
                    completion tracked: success=True                               
           DEBUG    [01:54:04] selector_events.py:64 DEBUG    selector_events.py:64
                    Using selector: KqueueSelector                                 
Result: 4

Telemetry metrics collected:
- Agent executions: 1
- Task completions: 1
- Errors: 0
- Session ID: 7cb1e713629a7b0f

To disable telemetry, set any of these environment variables:
- PRAISONAI_TELEMETRY_DISABLED=true
- DO_NOT_TRACK=true
           DEBUG    [01:54:04] telemetry.py:207 DEBUG Telemetry    telemetry.py:207
                    flush: {'enabled': True, 'session_id':                         
                    '7cb1e713629a7b0f', 'metrics':                                 
                    {'agent_executions': 1, 'task_completions': 1,                 
                    'tool_calls': 0, 'errors': 0}, 'environment':                  
                    {'python_version': '3.13.3', 'os_type':                        
                    'Darwin', 'framework_version': 'unknown'}}    

[claude]

Claude finished @MervinPraison's task in 2m 5s —— View job • claude/issue-579-20250607_005524 • Create PR ➔

---

Telemetry PostHog Issue - RESOLVED ✅

Looking into why telemetry data is not being posted to PostHog despite being collected locally.

Todo List:

• Examine telemetry implementation code structure
• Check telemetry_minimal.py script
• Identify PostHog integration logic
• Diagnose the root cause
• Provide solution recommendations
• Fix the missing dependency issue
• Create diagnostic tool

Root Cause: Missing PostHog Dependency

The issue was that PostHog was not installed in your environment, even though telemetry data was being collected locally.

What was happening:

1. The telemetry code in src/praisonai-agents/praisonaiagents/telemetry/telemetry.py:16-21 tries to import PostHog
2. If PostHog is not installed, it fails silently (POSTHOG_AVAILABLE = False)
3. Telemetry metrics are collected locally (as shown in your logs)
4. During flush(), no data is sent to PostHog because _posthog = None

The Problem:

• posthog>=3.0.0 was listed in pyproject.toml as a core dependency
• But it was missing from requirements.txt
• If you installed with pip install -r requirements.txt, PostHog wouldn't be installed

Solutions:

Option 1: Install PostHog directly

pip install posthog>=3.0.0

Option 2: Reinstall with telemetry extras

pip install praisonaiagents[telemetry]

Option 3: Use the diagnostic script to verify

python debug_posthog.py

Changes Made:

• ✅ Added posthog>=3.0.0 to requirements.txt to fix the dependency issue
• ✅ Created debug_posthog.py diagnostic script to help troubleshoot PostHog issues

After installing PostHog (pip install posthog>=3.0.0), run your telemetry_minimal.py script again and you should see telemetry data being sent to PostHog.

---