What to Install

Install this module when you need background/async generation for long-running tasks.

odoo-bin -d your_db -i llm_generate_job

Auto-Installed Dependencies

These are pulled in automatically:

• llm_generate (generation API)
• llm_thread (chat interface)
• llm_tool (tool framework)
• llm (core infrastructure)

When to Use This Module

Common Setups

Generation Job Management

• Job Lifecycle: Complete job lifecycle management from creation to completion
• Status Tracking: Real-time job status (draft, queued, running, completed, failed, cancelled)
• Retry Logic: Automatic and manual retry capabilities for failed jobs
• Error Handling: Comprehensive error tracking and reporting

Queue Management

• Provider-specific Queues: Each LLM provider has its own dedicated queue
• Concurrent Job Control: Configurable maximum concurrent jobs per provider
• Queue Health Monitoring: Real-time queue health indicators
• Performance Metrics: Queue performance analytics and success rates

Basic Usage

# Direct generation (backward compatible)
thread = self.env['llm.thread'].browse(thread_id)
for update in thread.generate_response("Hello", use_queue=False):
    print(update)

# Queued generation
for update in thread.generate_response("Hello", use_queue=True):
    print(update)

# Auto-detection (recommended)
for update in thread.generate_response("Hello"):
    # Automatically chooses based on provider capabilities
    print(update)

Job Management

# Create a job
job = self.env['llm.generation.job'].create({
    'thread_id': thread_id,
    'provider_id': provider_id,
    'model_id': model_id,
    'generation_inputs': {'prompt': 'Hello world'},
})

# Queue and start the job
job.action_queue()

# Monitor job status
while job.state in ['queued', 'running']:
    status = job.check_status()
    print(f"Job {job.id} is {job.state}")

Models

Queue Health

Queues are automatically monitored for:

• Healthy: Normal operation
• Warning: High load but functioning
• Critical: Overloaded or failing
• Disabled: Manually disabled

Performance Metrics

• Average Queue Time: Time jobs spend waiting
• Average Processing Time: Time jobs spend processing
• Success Rate: Percentage of successful jobs
• Throughput: Jobs processed per time period

Cron Jobs

• Process Queues: Automatically process pending jobs (every minute)
• Check Job Status: Update running job statuses (every 30 seconds)
• Auto-retry Failed Jobs: Retry eligible failed jobs (every 5 minutes)
• Cleanup Old Jobs: Remove old completed jobs (daily)

Module Information

• Name: LLM Generate Job
• Version: 18.0.1.0.0
• Category: Productivity
• License: LGPL-3
• Dependencies: llm_thread, llm_tool
• Author: Apexive Solutions LLC