fix(deploy): resolve releases without the GitHub API in install.sh

[shiv-tyagi]

What

curl ... install.sh | bash failed with No releases found. Two causes:

1. The unauthenticated GitHub API is rate-limited to 60 req/hr per IP; shared egress IPs hit 403, which surfaced as the misleading error.
2. /releases/latest ignores pre-releases, and every Spur release is a pre-release, so it returns 404 even when authenticated.

Fix

Drop the GitHub API. Resolve the latest tag from the releases.atom feed (newest-first, first non-nightly tag) and the nightly asset from expanded_assets. Both are served by github.com with no rate limit and include pre-releases.

Testing

Ran install.sh for both latest (resolves v0.3.0) and nightly from a previously rate-limited IP. Both download, verify checksum, and install cleanly.

[Copilot]

Pull request overview

This PR broadens beyond the stated deploy/install fix by (a) changing install.sh release resolution to avoid GitHub API rate limits and pre-release handling issues, and (b) introducing node drain/removal and agent deregistration plumbing across the controller/agent/proto/CLI plus new E2E + unit coverage for node-down eviction behavior.

Changes:

• Update install.sh to resolve latest via releases.atom and resolve nightly assets via expanded_assets (no GitHub API usage).
• Add node drain/remove and agent self-deregistration APIs (proto + spurctld server + spur-cli), including eviction/job finalization on node down/remove and a configurable heartbeat timeout.
• Add new E2E tests and expand unit tests around eviction and node state transitions.

Reviewed changes

Copilot reviewed 1 out of 1 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
tests/native_host/e2e/test_deregistration.py	Adds E2E coverage for drain/remove, forced eviction, agent SIGTERM deregistration, and down-node job failure.
tests/native_host/e2e/conftest.py	Introduces cluster_config_overrides fixture and threads overrides into cluster deployment fixtures.
tests/native_host/e2e/cluster.py	Broadens stale-agent cleanup to also kill any process holding the agent port.
proto/slurm.proto	Adds controller RPCs for draining/removing nodes and agent-triggered deregistration.
install.sh	Removes GitHub API calls for resolving latest/nightly release artifacts.
crates/spurd/src/reporter.rs	Adds agent-side deregistration RPC call into the controller.
crates/spurd/src/main.rs	Handles SIGTERM to attempt deregistration before process exit.
crates/spurctld/src/server.rs	Implements new gRPC endpoints for drain/remove/deregister flows and forwards to leader when needed.
crates/spurctld/src/raft.rs	Changes Raft client response from single finalized job to a list of finalized jobs.
crates/spurctld/src/main.rs	Makes node heartbeat timeout configurable and triggers cancel + finalize steps for evictions.
crates/spurctld/src/cluster.rs	Implements drain/remove operations, eviction logic, and multi-job finalization side effects.
crates/spur-core/src/wal.rs	Adds WAL op for node removal + serde round-trip tests.
crates/spur-core/src/job.rs	Allows Completing -> NodeFail transition.
crates/spur-core/src/config.rs	Adds controller.heartbeat_timeout_secs configuration and tests.
crates/spur-cli/src/node.rs	Adds spur node drain and spur node remove commands.
.github/workflows/ci.yml	Changes Trivy scan severities and makes scans non-blocking via exit-code: "0".

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #383      +/-   ##
==========================================
- Coverage   68.21%   68.21%   -0.00%     
==========================================
  Files         134      134              
  Lines       36088    36098      +10     
==========================================
+ Hits        24615    24621       +6     
- Misses      11473    11477       +4     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.