[misc] Pre-release cleanup

[chus-chus]

Updated auto-managed server support and other minor additions

Summary by CodeRabbit

Release Notes

• Chores

  • Removed Vajra server type support from all configuration and orchestration components.

• Bug Fixes

  • Improved resource manager timeout handling with more accurate timing measurement and optimized GPU allocation logic for better resource availability.

• Documentation

  • Updated supported servers documentation to reflect current server type options.

[coderabbitai]

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Free

Run ID: 62033f92-1fa2-4a1f-b311-4d0e4d6cdfdc

📥 Commits

Reviewing files that changed from the base of the PR and between f26a28f and 4a052f6.

📒 Files selected for processing (3)

• tests/unit/orchestration/test_engine_server_commands.py
• tests/unit/orchestration/test_resource_manager.py
• veeksha/orchestration/resource_manager.py

🚧 Files skipped from review as they are similar to previous changes (1)

• tests/unit/orchestration/test_engine_server_commands.py

---

📝 Walkthrough

Walkthrough

Removes Vajra server support across code, tests, and docs: enum member, config dataclass, server manager implementation and registry entry, related tests, and documentation references; also updates resource allocation timing and polling behavior.

Changes

Cohort / File(s)	Summary
Type Definitions & Config veeksha/types/__init__.py, veeksha/config/server.py	Removes VAJRA enum member (reassigns SGLANG value) and deletes VajraServerConfig dataclass (and its get_type() override).
Orchestration: Vajra removal veeksha/orchestration/vajra_server.py, veeksha/orchestration/__init__.py, veeksha/orchestration/registry.py	Deletes VajraServerManager class and module logger; removes import/export of VajraServerManager and registry registration mapping for ServerType.VAJRA.
Resource manager runtime logic veeksha/orchestration/resource_manager.py	Switches to time.monotonic() for timeouts, adjusts allocation retry conditions to check total/free GPUs, reuses computed free/total values in logs, and makes sleep adaptive to remaining timeout.
Tests: orchestration & commands tests/unit/orchestration/test_benchmark_orchestrator.py, tests/unit/orchestration/test_engine_server_commands.py	Removes Vajra-specific tests and imports; updates YAML-driven config test to use type: vllm and create_class_from_dict(BaseServerConfig, ...) instead of engine-branching; adjusts imports accordingly.
Tests: resource manager tests/unit/orchestration/test_resource_manager.py	Adds time import and measures real elapsed time in test_wait_for_resources_timeout, asserting allocation is None and elapsed under threshold; adjusts poll interval usage.
Documentation docs/user_guide/server_management.rst	Removes Vajra from supported servers list and shortens inline server.type example to reference only sglang and vllm.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Poem

🐇 I nibbled code, then gave a twitch,

Vajra hopped off without a hitch,
VLLM and SgLang keep the race,
Quiet burrows, tidier place. 🥕

---

Note

🎁 Summarized by CodeRabbit Free

Your organization is on the Free plan. CodeRabbit will generate a high-level summary and a walkthrough for each pull request. For a comprehensive line-by-line review, please upgrade your subscription to CodeRabbit Pro by visiting https://app.coderabbit.ai/login.

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[github-actions]

	Tests	Passed ✅	Skipped	Failed
JUnit Test Report	262 ran	262 passed	0 skipped	0 failed

• View JUnit Test Report