Add orchestrator CLI command with REST API for task queue management by konard · Pull Request #1194 · link-assistant/hive-mind

konard · 2026-01-29T00:41:58Z

Summary

Implements issue #1193 - Add orchestrator CLI command for centralized task queue management.

New Features

orchestrator CLI command with configurable --port (default: 8080) and --hostname (default: 0.0.0.0)
REST API using LINO format (Links Notation - alternative to JSON):
- POST /api/v0/solve/enqueue - enqueue solve tasks with issue URL and options
- GET /api/v0/solve/queue - retrieve queue status and stats
- GET /api/v0/solve/task/:id - get details for a specific task
- GET /health - health check endpoint
--use-orchestrator option for solve, hive, and telegram-bot commands to delegate tasks to orchestrator
Upstream load balancing support (master/slave pattern) via --upstream option
Queue management with resource-aware throttling (RAM, CPU, disk, Claude/GitHub limits)
Separate queues per tool type (PR feat: separate queues for claude and agent tools #1167): Claude tasks never block agent tasks and vice versa

Files Added

src/orchestrator.mjs - Main CLI entry point
src/orchestrator.config.lib.mjs - CLI configuration/argument parsing
src/orchestrator-queue.lib.mjs - Queue management logic with tool-specific queues (aligned with telegram-solve-queue.lib.mjs from PR feat: separate queues for claude and agent tools #1167)
src/orchestrator-client.lib.mjs - Client library for communicating with orchestrator
dependencies/lino-rest-api/ - LINO REST API implementation (codec, middleware, app)
tests/orchestrator-*.test.mjs - Unit tests (29 tests)
tests/lino-codec.test.mjs - LINO codec tests

Files Modified

package.json - Added orchestrator binary
src/solve.config.lib.mjs - Added --use-orchestrator option
src/hive.config.lib.mjs - Added --use-orchestrator option
src/telegram-bot.mjs - Added --useOrchestrator option
eslint.config.mjs - Extended to include dependencies folder
tests/solve-queue.test.mjs - Fixed test expecting incorrect threshold value (0.85 → 0.75)

Architecture: Separate Queues per Tool (Issue #1159)

// Instead of single queue:
this.queue = [];                    // Single queue for all tools

// Now separate queues:
this.queues = {                     // Separate queues per tool
  claude: [],
  agent: [],
};
this.lastStartTimeByTool = {        // Independent timing per tool
  claude: null,
  agent: null,
};

Key changes applied from PR #1167:

getToolQueue(tool) method for accessing tool-specific queues
queue getter for backwards compatibility (returns combined queue)
getTotalQueueLength() and getProcessingCountByTool() methods
findStartableItems() to return startable items from each tool queue
Tool-specific canStartCommand() with independent lastStartTimeByTool
Agent tasks skip Claude API limits (5-hour session, weekly limits)

Test Plan

Run npm test - All tests pass (70 main + 18 limits + 63 usage = 151 tests)
Run node --test tests/orchestrator*.test.mjs - 29 tests pass
ESLint passes with no errors
CI checks pass
Start orchestrator: ./src/orchestrator.mjs --port 8080 --verbose
Test health endpoint: curl http://localhost:8080/health
Test queue endpoint: curl http://localhost:8080/api/v0/solve/queue
Test enqueue: curl -X POST http://localhost:8080/api/v0/solve/enqueue -H "Content-Type: text/lino" -d "issueUrl: https://github.com/owner/repo/issues/1"

Fixes #1193

🤖 Generated with Claude Code

Adding CLAUDE.md with task information for AI processing. This file will be removed when the task is complete. Issue: #1193

Implements issue #1193 with the following features: - New `orchestrator` CLI command with configurable --port and --hostname - REST API using lino-rest-api (Links Notation format): - POST /api/v0/solve/enqueue - enqueue solve tasks - GET /api/v0/solve/queue - retrieve queue status - GET /api/v0/solve/task/:id - get task details - GET /health - health check endpoint - New --use-orchestrator option for solve, hive, and telegram-bot commands - Support for upstream orchestrator load balancing (master/slave pattern) - Queue management with resource-aware throttling (RAM, CPU, disk, Claude/GitHub limits) - Unit tests for orchestrator-queue, orchestrator-client, and lino-codec Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Reduced verbose multi-line comments explaining configuration priority to a single concise line. This keeps the file under the 1500 line limit after adding the --useOrchestrator option. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

This reverts commit ade14bc.

konard · 2026-01-29T01:07:23Z

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

Public pricing estimate: $8.483700 USD
Calculated by Anthropic: $7.070713 USD
Difference: $-1.412986 (-16.66%)
📎 Log file uploaded as Gist (1458KB)
🔗 View complete solution draft log

Now working session is ended, feel free to review and add any feedback on the solution draft.

konard · 2026-01-29T01:11:28Z

Also make sure to apply changes from #1167.

And ensure all changes are correct, consistent and fully meet all discussed requirements (check issue description and all comments in issue and in pull request).

konard · 2026-01-29T10:20:12Z

🤖 AI Work Session Started

Starting automated work session at 2026-01-29T10:20:10.110Z

The PR has been converted to draft mode while work is in progress.

This comment marks the beginning of an AI work session. Please wait working session to finish, and provide your feedback.

Updates orchestrator-queue.lib.mjs with the same tool-specific queue implementation from PR #1167 (telegram-solve-queue): - Separate queues per tool type (claude, agent) instead of single queue - Tool-specific timing (lastStartTimeByTool) for independent processing - canStartCommand() and checkApiLimits() accept tool parameter - Skip Claude API limits when tool is 'agent' - New methods: getToolQueue(), getTotalQueueLength(), getProcessingCountByTool() - New methods: findStartableItems(), findStartableItem() - Updated consumer loop to check each tool queue independently - Add updateAllWaitingItems() for tool-specific waiting reasons This ensures: - Claude tasks never block agent tasks (and vice versa) - Each tool queue maintains FIFO order - Each tool has independent rate limiting Also adds comprehensive tests for the new functionality: - Tool-specific queue tests - Cross-queue findByUrl and findById tests - Per-tool cancellation tests - Stats with per-tool breakdown tests - Tool-specific limit handling tests See: #1159 Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

The test was asserting 0.85 but the actual QUEUE_CONFIG uses 0.75. Updated the test to match the current configuration value. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Use Object.values() instead of Object.entries() when we only need the queue arrays, not the tool keys. The tool is already included in the item's toJSON() output. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

konard · 2026-01-29T10:41:01Z

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

Public pricing estimate: $7.641584 USD
Calculated by Anthropic: $5.969452 USD
Difference: $-1.672132 (-21.88%)
📎 Log file uploaded as Gist (1555KB)
🔗 View complete solution draft log

Now working session is ended, feel free to review and add any feedback on the solution draft.

konard · 2026-01-29T10:41:08Z

🔄 Auto-restart 1/3

Detected uncommitted changes from previous run. Starting new session to review and commit them.

Uncommitted files:

?? issue_1193.txt
?? issue_comments.json
?? pr_1167_diff.txt
?? pr_1167_view.txt
?? pr_conversation_comments.json
?? pr_review_comments.json
?? pr_view.txt

Auto-restart will stop after changes are committed or after 2 more iterations. Please wait until working session will end and give your feedback.

konard · 2026-01-29T10:45:18Z

🔄 Auto-restart 1/3 Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

Public pricing estimate: $2.337334 USD
Calculated by Anthropic: $1.013733 USD
Difference: $-1.323601 (-56.63%)
📎 Log file uploaded as Gist (1955KB)
🔗 View complete solution draft log

Now working session is ended, feel free to review and add any feedback on the solution draft.

konard · 2026-01-29T10:45:26Z

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

Public pricing estimate: $2.337334 USD
Calculated by Anthropic: $1.013733 USD
Difference: $-1.323601 (-56.63%)
📎 Log file uploaded as Gist (1959KB)
🔗 View complete solution draft log

Now working session is ended, feel free to review and add any feedback on the solution draft.

Initial commit with task details

ade14bc

Adding CLAUDE.md with task information for AI processing. This file will be removed when the task is complete. Issue: #1193

konard self-assigned this Jan 29, 2026

konard changed the title ~~[WIP] Add orchestartor CLI command~~ Add orchestrator CLI command with REST API for task queue management Jan 29, 2026

konard marked this pull request as ready for review January 29, 2026 00:59

konard and others added 2 commits January 29, 2026 02:01

Revert "Initial commit with task details"

4920e77

This reverts commit ade14bc.

konard marked this pull request as draft January 29, 2026 10:20

konard and others added 4 commits January 29, 2026 11:21

Merge remote-tracking branch 'origin/main' into issue-1193-172b71650d82

b262a92

Fix CLAUDE_5_HOUR_SESSION_THRESHOLD test to match actual value (0.75)

09ee549

The test was asserting 0.85 but the actual QUEUE_CONFIG uses 0.75. Updated the test to match the current configuration value. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Fix ESLint no-unused-vars errors in orchestrator-queue.lib.mjs

f01f4b7

Use Object.values() instead of Object.entries() when we only need the queue arrays, not the tool keys. The tool is already included in the item's toJSON() output. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

konard marked this pull request as ready for review January 29, 2026 10:40

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add orchestrator CLI command with REST API for task queue management#1194

Add orchestrator CLI command with REST API for task queue management#1194
konard wants to merge 8 commits intomainfrom
issue-1193-172b71650d82

konard commented Jan 29, 2026 •

edited

Loading

Uh oh!

konard commented Jan 29, 2026

Uh oh!

konard commented Jan 29, 2026 •

edited

Loading

Uh oh!

konard commented Jan 29, 2026

Uh oh!

konard commented Jan 29, 2026

Uh oh!

konard commented Jan 29, 2026

Uh oh!

konard commented Jan 29, 2026

Uh oh!

konard commented Jan 29, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

konard commented Jan 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

New Features

Files Added

Files Modified

Architecture: Separate Queues per Tool (Issue #1159)

Test Plan

Uh oh!

konard commented Jan 29, 2026

🤖 Solution Draft Log

Uh oh!

konard commented Jan 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

konard commented Jan 29, 2026

Uh oh!

konard commented Jan 29, 2026

🤖 Solution Draft Log

Uh oh!

konard commented Jan 29, 2026

🔄 Auto-restart 1/3

Uh oh!

konard commented Jan 29, 2026

🔄 Auto-restart 1/3 Log

Uh oh!

konard commented Jan 29, 2026

🤖 Solution Draft Log

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

konard commented Jan 29, 2026 •

edited

Loading

konard commented Jan 29, 2026 •

edited

Loading