mirror of
https://github.com/headroomlabs-ai/headroom.git
synced 2026-08-27 14:17:10 -04:00
## Description Headroom fails silently: a client not routed through the proxy (or a proxy running stale code) keeps working — it just stops saving tokens. State that determines whether you are actually saving lives in five places nothing reconciles. `headroom doctor` correlates them in one command (the diagnostic idiom of `claude doctor` / `pnpm doctor`, and the repo's own `headroom tools doctor`). Closes # <!-- no tracked issue; setup-diagnosis gap found this session --> ## Type of Change - [ ] Bug fix (non-breaking change that fixes an issue) - [x] New feature (non-breaking change that adds functionality) - [ ] Breaking change (fix or feature that would cause existing functionality to change) - [ ] Documentation update - [ ] Performance improvement - [ ] Code refactoring (no functional changes) ## Changes Made - `headroom/cli/doctor.py`: new command with 8 pure checks (proxy liveness, version drift, claude/codex routing, shell env, savings flow, budget, deployments); exit codes 0/1/2; `--json`; `--port`/`HEADROOM_PORT`. - `headroom/proxy/cost.py`: expose `budget_limit_usd`/`budget_period` in `CostTracker.stats()` so the budget check can read it (older proxies degrade to a warning). - `headroom/cli/main.py`: register the command. - `tests/test_cli_doctor.py`: 41 tests, zero network (probed payloads / paths / env injected). ## Testing - [x] Unit tests pass (`pytest`) - [x] Linting passes (`ruff check .`) - [x] Type checking passes (`mypy headroom`) - [x] New tests added for new functionality - [x] Manual testing performed ### Test Output ```text $ pytest tests/test_cli_doctor.py -q 41 passed ``` ## Real Behavior Proof - Environment: local macOS, Python 3.11, against a real proxy running for 3 days, branch `feat/doctor-command`. - Exact command / steps: `headroom doctor` (live), plus `pytest tests/test_cli_doctor.py -q`. - Observed result: Correctly flagged real version drift (proxy 0.25.0 vs installed 0.26.0), an unrouted claude client, and a shell `OPENAI_BASE_URL` pointed at a non-Headroom gateway; savings check showed 17.6M tokens / $7.82 saved; exit code 1 (warnings). - Not tested: Windows path handling for client config files (logic is OS-agnostic via pathlib). ## Review Readiness - [x] I have performed a self-review - [x] This PR is ready for human review ## Checklist - [x] My code follows the project's style guidelines - [x] I have performed a self-review of my code - [x] I have commented my code, particularly in hard-to-understand areas - [ ] I have made corresponding changes to the documentation - [x] My changes generate no new warnings - [x] I have added tests that prove my fix is effective or that my feature works - [x] New and existing unit tests pass locally with my changes - [ ] I have updated the CHANGELOG.md if applicable ## Screenshots (if applicable) Terminal output of `headroom doctor` (rich table) can be attached; the rendered table is reproduced in the live-proof bullet above. ## Additional Notes Branched fresh from main. The budget check connects to the enforcement fix in #885. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com> Co-authored-by: JD Davis <mxjerrett@gmail.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
349 lines
12 KiB
Python
349 lines
12 KiB
Python
"""Tests for `headroom doctor`."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import json
|
|
from dataclasses import dataclass
|
|
|
|
import pytest
|
|
from click.testing import CliRunner
|
|
|
|
import headroom.cli.doctor as doctor_mod
|
|
from headroom.cli.doctor import (
|
|
FAIL,
|
|
PASS,
|
|
SKIP,
|
|
WARN,
|
|
check_budget,
|
|
check_claude_routing,
|
|
check_codex_routing,
|
|
check_deployments,
|
|
check_proxy_liveness,
|
|
check_savings,
|
|
check_shell_env,
|
|
check_version_drift,
|
|
)
|
|
from headroom.cli.main import main
|
|
|
|
LIVEZ_OK = {
|
|
"service": "headroom-proxy",
|
|
"status": "healthy",
|
|
"alive": True,
|
|
"version": "0.26.0",
|
|
"uptime_seconds": 260135.0,
|
|
}
|
|
|
|
STATS_OK = {
|
|
"persistent_savings": {
|
|
"lifetime": {"tokens_saved": 17_583_102, "compression_savings_usd": 7.81701},
|
|
"display_session": {"last_activity_at": "2026-06-12T12:00:00Z"},
|
|
},
|
|
"cost": {"budget_limit_usd": 10.0, "budget_period": "daily"},
|
|
}
|
|
|
|
|
|
class TestProxyLiveness:
|
|
def test_down_is_fail_with_hint(self):
|
|
result = check_proxy_liveness(None, "http://127.0.0.1:8787")
|
|
assert result.status == FAIL
|
|
assert "headroom proxy" in (result.hint or "")
|
|
|
|
def test_up_mentions_version_and_uptime(self):
|
|
result = check_proxy_liveness(LIVEZ_OK, "http://127.0.0.1:8787")
|
|
assert result.status == PASS
|
|
assert "v0.26.0" in result.summary
|
|
assert "3d" in result.summary
|
|
|
|
|
|
class TestVersionDrift:
|
|
def test_match_passes(self):
|
|
assert check_version_drift(LIVEZ_OK, "0.26.0").status == PASS
|
|
|
|
def test_mismatch_warns_with_restart_hint(self):
|
|
result = check_version_drift(LIVEZ_OK, "0.27.0")
|
|
assert result.status == WARN
|
|
assert "drift" in result.summary
|
|
assert "restart" in (result.hint or "")
|
|
|
|
def test_proxy_down_skips(self):
|
|
assert check_version_drift(None, "0.26.0").status == SKIP
|
|
|
|
def test_unknown_version_warns(self):
|
|
assert check_version_drift({"version": "unknown"}, "0.26.0").status == WARN
|
|
assert check_version_drift(LIVEZ_OK, "unknown").status == WARN
|
|
|
|
|
|
class TestClaudeRouting:
|
|
def test_missing_file_warns(self, tmp_path):
|
|
result = check_claude_routing(tmp_path / "settings.json", 8787)
|
|
assert result.status == WARN
|
|
assert "wrap claude" in (result.hint or "")
|
|
|
|
def test_malformed_json_warns(self, tmp_path):
|
|
path = tmp_path / "settings.json"
|
|
path.write_text("{not json", encoding="utf-8")
|
|
assert check_claude_routing(path, 8787).status == WARN
|
|
|
|
def test_no_env_key_warns(self, tmp_path):
|
|
path = tmp_path / "settings.json"
|
|
path.write_text(json.dumps({"env": {}}), encoding="utf-8")
|
|
assert check_claude_routing(path, 8787).status == WARN
|
|
|
|
def test_correct_url_passes(self, tmp_path):
|
|
path = tmp_path / "settings.json"
|
|
path.write_text(
|
|
json.dumps({"env": {"ANTHROPIC_BASE_URL": "http://127.0.0.1:8787"}}),
|
|
encoding="utf-8",
|
|
)
|
|
assert check_claude_routing(path, 8787).status == PASS
|
|
|
|
def test_port_mismatch_warns(self, tmp_path):
|
|
path = tmp_path / "settings.json"
|
|
path.write_text(
|
|
json.dumps({"env": {"ANTHROPIC_BASE_URL": "http://127.0.0.1:8788"}}),
|
|
encoding="utf-8",
|
|
)
|
|
result = check_claude_routing(path, 8787)
|
|
assert result.status == WARN
|
|
assert "8788" in result.summary
|
|
|
|
def test_non_headroom_url_warns(self, tmp_path):
|
|
path = tmp_path / "settings.json"
|
|
path.write_text(
|
|
json.dumps({"env": {"ANTHROPIC_BASE_URL": "https://gateway.corp.example/v1"}}),
|
|
encoding="utf-8",
|
|
)
|
|
result = check_claude_routing(path, 8787)
|
|
assert result.status == WARN
|
|
assert "gateway.corp.example" in result.summary
|
|
|
|
|
|
class TestCodexRouting:
|
|
def test_missing_file_warns(self, tmp_path):
|
|
assert check_codex_routing(tmp_path / "config.toml", 8787).status == WARN
|
|
|
|
def test_marker_block_right_port_passes(self, tmp_path):
|
|
path = tmp_path / "config.toml"
|
|
path.write_text(
|
|
'model_provider = "headroom"\n'
|
|
"[model_providers.headroom]\n"
|
|
'base_url = "http://127.0.0.1:8787/v1"\n',
|
|
encoding="utf-8",
|
|
)
|
|
assert check_codex_routing(path, 8787).status == PASS
|
|
|
|
def test_port_mismatch_warns(self, tmp_path):
|
|
path = tmp_path / "config.toml"
|
|
path.write_text(
|
|
'[model_providers.headroom]\nbase_url = "http://127.0.0.1:9999/v1"\n',
|
|
encoding="utf-8",
|
|
)
|
|
result = check_codex_routing(path, 8787)
|
|
assert result.status == WARN
|
|
assert "9999" in result.summary
|
|
|
|
def test_no_marker_warns(self, tmp_path):
|
|
path = tmp_path / "config.toml"
|
|
path.write_text('model = "gpt-5"\n', encoding="utf-8")
|
|
assert check_codex_routing(path, 8787).status == WARN
|
|
|
|
def test_garbage_bytes_warn_not_crash(self, tmp_path):
|
|
path = tmp_path / "config.toml"
|
|
path.write_bytes(b"\xff\xfe garbage \x00")
|
|
assert check_codex_routing(path, 8787).status == WARN
|
|
|
|
|
|
class TestShellEnv:
|
|
def test_unset_warns(self):
|
|
result = check_shell_env({}, 8787)
|
|
assert result.status == WARN
|
|
assert "bypasses" in result.summary
|
|
|
|
def test_matching_anthropic_url_passes(self):
|
|
env = {"ANTHROPIC_BASE_URL": "http://127.0.0.1:8787"}
|
|
assert check_shell_env(env, 8787).status == PASS
|
|
|
|
def test_localhost_also_passes(self):
|
|
env = {"OPENAI_BASE_URL": "http://localhost:8787/v1"}
|
|
assert check_shell_env(env, 8787).status == PASS
|
|
|
|
def test_other_url_warns(self):
|
|
env = {"ANTHROPIC_BASE_URL": "https://api.anthropic.com"}
|
|
assert check_shell_env(env, 8787).status == WARN
|
|
|
|
|
|
class TestSavings:
|
|
def test_from_stats_passes_with_totals(self, tmp_path):
|
|
result = check_savings(STATS_OK, tmp_path / "missing.json")
|
|
assert result.status == PASS
|
|
assert "17,583,102" in result.summary
|
|
assert "$7.82" in result.summary
|
|
|
|
def test_falls_back_to_file_when_proxy_down(self, tmp_path):
|
|
savings_file = tmp_path / "proxy_savings.json"
|
|
savings_file.write_text(
|
|
json.dumps(
|
|
{
|
|
"lifetime": {"tokens_saved": 500, "compression_savings_usd": 0.01},
|
|
"display_session": {"last_activity_at": "2026-06-12T11:00:00Z"},
|
|
}
|
|
),
|
|
encoding="utf-8",
|
|
)
|
|
result = check_savings(None, savings_file)
|
|
assert result.status == PASS
|
|
assert "500" in result.summary
|
|
assert str(savings_file) in result.summary
|
|
|
|
def test_no_data_warns(self, tmp_path):
|
|
assert check_savings(None, tmp_path / "missing.json").status == WARN
|
|
|
|
def test_zero_tokens_warns(self, tmp_path):
|
|
stats = {"persistent_savings": {"lifetime": {"tokens_saved": 0}}}
|
|
assert check_savings(stats, tmp_path / "missing.json").status == WARN
|
|
|
|
|
|
class TestBudget:
|
|
def test_proxy_down_skips(self):
|
|
assert check_budget(None).status == SKIP
|
|
|
|
def test_cost_tracking_disabled_warns(self):
|
|
assert check_budget({"cost": None}).status == WARN
|
|
|
|
def test_old_proxy_without_keys_warns(self):
|
|
result = check_budget({"cost": {"savings_usd": 1.0}})
|
|
assert result.status == WARN
|
|
assert "older version" in result.summary
|
|
|
|
def test_unset_budget_warns_with_hint(self):
|
|
result = check_budget({"cost": {"budget_limit_usd": None}})
|
|
assert result.status == WARN
|
|
assert "--budget" in (result.hint or "")
|
|
|
|
def test_configured_budget_passes(self):
|
|
result = check_budget(STATS_OK)
|
|
assert result.status == PASS
|
|
assert "$10.0/daily" in result.summary
|
|
|
|
|
|
@dataclass
|
|
class _FakeManifest:
|
|
profile: str
|
|
health_url: str
|
|
|
|
|
|
class TestDeployments:
|
|
def test_no_manifests_omits_section(self):
|
|
assert check_deployments([]) is None
|
|
|
|
def test_all_healthy_passes(self):
|
|
manifests = [_FakeManifest("default", "http://127.0.0.1:8787/readyz")]
|
|
result = check_deployments(manifests, probe=lambda url: {"ready": True})
|
|
assert result is not None and result.status == PASS
|
|
|
|
def test_unhealthy_fails_naming_profile(self):
|
|
manifests = [_FakeManifest("prod", "http://127.0.0.1:9999/readyz")]
|
|
result = check_deployments(manifests, probe=lambda url: None)
|
|
assert result is not None and result.status == FAIL
|
|
assert "prod" in result.summary
|
|
|
|
|
|
class TestDoctorCommand:
|
|
@pytest.fixture
|
|
def runner(self):
|
|
return CliRunner()
|
|
|
|
@pytest.fixture
|
|
def isolated(self, tmp_path, monkeypatch):
|
|
"""Point all filesystem/network surfaces at controlled fakes."""
|
|
monkeypatch.setattr(doctor_mod, "claude_settings_path", lambda: tmp_path / "settings.json")
|
|
monkeypatch.setattr(doctor_mod, "codex_config_path", lambda: tmp_path / "config.toml")
|
|
monkeypatch.setattr(doctor_mod, "savings_path", lambda: tmp_path / "savings.json")
|
|
monkeypatch.setattr(doctor_mod, "list_manifests", lambda: [])
|
|
for var in ("ANTHROPIC_BASE_URL", "OPENAI_BASE_URL", "HEADROOM_PORT"):
|
|
monkeypatch.delenv(var, raising=False)
|
|
return tmp_path
|
|
|
|
def _probe(self, livez, stats):
|
|
def fake_probe(url, timeout=2.0):
|
|
if url.endswith("/livez"):
|
|
return livez
|
|
if url.endswith("/stats"):
|
|
return stats
|
|
return None
|
|
|
|
return fake_probe
|
|
|
|
def test_proxy_down_exits_2(self, runner, isolated, monkeypatch):
|
|
monkeypatch.setattr(doctor_mod, "probe_json", self._probe(None, None))
|
|
result = runner.invoke(main, ["doctor"])
|
|
assert result.exit_code == 2
|
|
assert "not reachable" in result.output
|
|
|
|
def test_warnings_only_exits_1(self, runner, isolated, monkeypatch):
|
|
monkeypatch.setattr(doctor_mod, "probe_json", self._probe(LIVEZ_OK, STATS_OK))
|
|
monkeypatch.setattr(doctor_mod, "get_version", lambda: "0.26.0")
|
|
# proxy healthy, but clients unwrapped + shell env unset -> warns
|
|
result = runner.invoke(main, ["doctor"])
|
|
assert result.exit_code == 1
|
|
|
|
def test_all_pass_exits_0(self, runner, isolated, monkeypatch):
|
|
monkeypatch.setattr(doctor_mod, "probe_json", self._probe(LIVEZ_OK, STATS_OK))
|
|
monkeypatch.setattr(doctor_mod, "get_version", lambda: "0.26.0")
|
|
(isolated / "settings.json").write_text(
|
|
json.dumps({"env": {"ANTHROPIC_BASE_URL": "http://127.0.0.1:8787"}}),
|
|
encoding="utf-8",
|
|
)
|
|
(isolated / "config.toml").write_text(
|
|
'[model_providers.headroom]\nbase_url = "http://127.0.0.1:8787/v1"\n',
|
|
encoding="utf-8",
|
|
)
|
|
result = runner.invoke(
|
|
main, ["doctor"], env={"ANTHROPIC_BASE_URL": "http://127.0.0.1:8787"}
|
|
)
|
|
assert result.exit_code == 0, result.output
|
|
assert "all checks passed" in result.output
|
|
|
|
def test_json_output_parses(self, runner, isolated, monkeypatch):
|
|
monkeypatch.setattr(doctor_mod, "probe_json", self._probe(LIVEZ_OK, STATS_OK))
|
|
result = runner.invoke(main, ["doctor", "--json"])
|
|
payload = json.loads(result.output)
|
|
assert payload["port"] == 8787
|
|
assert {c["name"] for c in payload["checks"]} >= {"proxy", "version", "budget"}
|
|
assert all(c["status"] in ("pass", "warn", "fail", "skip") for c in payload["checks"])
|
|
|
|
def test_port_option_changes_probe_url(self, runner, isolated, monkeypatch):
|
|
seen: list[str] = []
|
|
|
|
def recording_probe(url, timeout=2.0):
|
|
seen.append(url)
|
|
return None
|
|
|
|
monkeypatch.setattr(doctor_mod, "probe_json", recording_probe)
|
|
runner.invoke(main, ["doctor", "--port", "9999"])
|
|
assert "http://127.0.0.1:9999/livez" in seen
|
|
|
|
def test_port_env_var_respected(self, runner, isolated, monkeypatch):
|
|
seen: list[str] = []
|
|
|
|
def recording_probe(url, timeout=2.0):
|
|
seen.append(url)
|
|
return None
|
|
|
|
monkeypatch.setattr(doctor_mod, "probe_json", recording_probe)
|
|
runner.invoke(main, ["doctor"], env={"HEADROOM_PORT": "9999"})
|
|
assert "http://127.0.0.1:9999/livez" in seen
|
|
|
|
|
|
class TestCostTrackerBudgetKeys:
|
|
def test_stats_exposes_budget_config(self):
|
|
from headroom.proxy.cost import CostTracker
|
|
|
|
stats = CostTracker(budget_limit_usd=5.0, budget_period="monthly").stats()
|
|
assert stats["budget_limit_usd"] == 5.0
|
|
assert stats["budget_period"] == "monthly"
|
|
|
|
def test_stats_budget_none_when_unset(self):
|
|
from headroom.proxy.cost import CostTracker
|
|
|
|
assert CostTracker().stats()["budget_limit_usd"] is None
|