Repository navigation
feat(OMN-1934): Add onex-infra-test integration test CLI - #261
Merged
jonahgabriel merged 12 commits intoFeb 8, 2026
Merged
Conversation
Scaffold the `onex-infra-test` CLI for end-to-end validation of the ONEX registration pipeline. Implements all 8 tasks from Phase 5: - P5.1: CLI scaffold using Click with Rich output - P5.2: `env up/down` wrapping docker-compose.e2e.yml with health wait - P5.3: `verify registry` checks Consul + PostgreSQL state - P5.4: `verify topics` validates ONEX 5-segment naming compliance - P5.5: `verify snapshots` checks compacted snapshot topic - P5.6: `verify idempotency` publishes N events, asserts <= 1 record - P5.7: `run --suite smoke` orchestrates happy-path verification - P5.8: `run --suite failure` tests runtime kill/restart recovery Also adds `introspect` command to publish test introspection events via rpk, and `run --suite idempotency` for duplicate detection. Entry point: `onex-infra-test` (registered in pyproject.toml)
- Remove unused imports (sys, asyncio) - Log exception details in bare except blocks (Consul, PostgreSQL) - Include exception type in all error messages for consistency - Document hardcoded test defaults as E2E-only
📝 WalkthroughWalkthroughAdds a new CLI package Changes
Sequence Diagram(s)sequenceDiagram
participant User as User/CLI
participant CLI as onex-infra-test CLI
participant Compose as Docker Compose
participant Kafka as Kafka (rpk)
participant DB as PostgreSQL
participant Consul as Consul
User->>CLI: invoke command (env / introspect / verify / run)
CLI->>Compose: start/stop services (compose-file, project-name)
Compose-->>CLI: status / health
CLI->>Kafka: publish introspection (rpk -> topic)
Kafka-->>CLI: publish result
CLI->>DB: poll registration_projections for node_id
DB-->>CLI: registration found / not found
CLI->>Consul: query KV / services
Consul-->>CLI: registry entries
CLI->>User: display PASS / WARN / FAIL
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing touches
🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 6
🤖 Fix all issues with AI agents
In `@src/omnibase_infra/cli/infra_test/introspect.py`:
- Around line 27-58: The function _build_introspection_payload currently ignores
the version parameter and hardcodes node_version; update
_build_introspection_payload to parse the version string argument (e.g.,
"1.2.3") into integers for major/minor/patch and populate "node_version" with
those values, handling missing segments by treating them as 0 and raising or
validating on non-numeric input; reference the function name
_build_introspection_payload and the "node_version" dict to locate where to
replace the hardcoded {"major":1,"minor":0,"patch":0} with the parsed values
from the version parameter.
In `@src/omnibase_infra/cli/infra_test/run_suite.py`:
- Around line 446-449: The inline message "Waiting for runtime health (30s)..."
in run_suite.py is inconsistent with the actual wait call; update the code so
the console.print and time.sleep match: either change time.sleep(10) to
time.sleep(30) to match the 30s message, or change the printed message to
"(10s)" to reflect the current time.sleep(10) behavior; modify the console.print
and/or time.sleep calls (the console.print(...) and time.sleep(...) lines)
accordingly so they are consistent.
- Around line 105-127: The polling loop may leak DB connections if an exception
occurs after psycopg2.connect() but before conn.close(); update the block that
calls psycopg2.connect(dsn, ...) and then cur = conn.cursor() / cur.execute(...)
to ensure cleanup by using context managers (with psycopg2.connect(...) as conn:
and with conn.cursor() as cur:) or a try/finally that always calls cur.close()
and conn.close(); specifically modify the connect/cursor/execute section that
uses dsn, node_id, timeout, console so that cur.close() and conn.close() are
guaranteed on error and successful paths.
- Around line 309-323: The current try/except around psycopg2.connect leaks
resources; update the block in run_suite to use context managers: open the
connection with "with psycopg2.connect(dsn, connect_timeout=5) as conn:" and the
cursor with "with conn.cursor() as cur:" around the cur.execute/fetchone logic
so conn and cur are closed automatically, preserve the existing exception
handling that logs the error (console.print with type(e).__name__ and e) and
then raises SystemExit(1).
- Around line 461-484: Replace the manual connect/cursor open/close in the
post-restart check with context managers to avoid leaking connections: use
psycopg2.connect(...) as conn and conn.cursor() as cur to run the SELECT against
registration_projections for entity_id node_a, fetchone and let the context
managers close the cursor and connection automatically; keep the existing
SystemExit and general Exception handling but remove explicit
cur.close()/conn.close() calls and ensure the same console prints and exit
behavior remain.
In `@tests/unit/cli/infra_test/test_introspect.py`:
- Around line 46-52: The test reveals _build_introspection_payload currently
ignores its version parameter and always sets node_version to
{"major":1,"minor":0,"patch":0}; update _build_introspection_payload to respect
the version argument (e.g., accept a string like "2.3.1" or a tuple/dict) by
parsing it and populating node_version accordingly (major/minor/patch ints) and
fallback to the default 1.0.0 when version is None or invalid, and then add a
unit test (besides test_has_version_semver) that passes a custom version (e.g.,
"2.3.1") and asserts node_version reflects 2, 3, 1; alternatively, if you prefer
not to support custom versions, remove the unused version parameter from
_build_introspection_payload and adjust tests to match the simplified signature.
🧹 Nitpick comments (10)
src/omnibase_infra/cli/infra_test/introspect.py (2)
113-128: Add a timeout to prevent indefinite hangs.The
subprocess.runcall lacks a timeout. Ifrpkhangs (e.g., broker unreachable without proper timeout handling in rpk itself), the CLI will block indefinitely.⏱️ Proposed fix to add timeout
result = subprocess.run( [ "rpk", "topic", "produce", topic, "--brokers", resolved_broker, "-k", str(nid), ], input=payload_json, capture_output=True, text=True, check=False, + timeout=30, )
96-98: Redundant environment variable resolution.The
--brokeroption already specifiesenvvar="KAFKA_BOOTSTRAP_SERVERS"(line 81), so Click will automatically resolve from the environment. The manualos.getenv("KAFKA_BOOTSTRAP_SERVERS")check is redundant.♻️ Simplified broker resolution
- resolved_broker: str = ( - broker or os.getenv("KAFKA_BOOTSTRAP_SERVERS") or "localhost:29092" - ) + resolved_broker: str = broker or "localhost:29092"src/omnibase_infra/cli/infra_test/verify.py (4)
29-31: Duplicate helper function.
_get_broker()is duplicated identically inrun_suite.py(lines 23-25). Consider extracting to a shared utility module to avoid drift.
96-101: Add timeout tosubprocess.runcalls.Multiple
subprocess.runcalls in this file (lines 96-101, 172-177, 192-209, 260-275) lack timeouts. Ifrpkhangs, the CLI will block indefinitely. Consider addingtimeout=30consistently.
385-408: Use context manager for database connections.The psycopg2 connection is not closed if an exception occurs between
connect()and the explicitclose()calls. Using a context manager ensures proper cleanup.🔒 Proposed fix using context manager
try: import psycopg2 - conn = psycopg2.connect(dsn, connect_timeout=5) - cur = conn.cursor() - - if node_id: - cur.execute( - "SELECT entity_id, current_state, node_type, updated_at " - "FROM registration_projections WHERE entity_id = %s", - (node_id,), - ) - else: - cur.execute( - "SELECT entity_id, current_state, node_type, updated_at " - "FROM registration_projections ORDER BY updated_at DESC LIMIT 20" - ) - - rows = cur.fetchall() - cur.close() - conn.close() + with psycopg2.connect(dsn, connect_timeout=5) as conn: + with conn.cursor() as cur: + if node_id: + cur.execute( + "SELECT entity_id, current_state, node_type, updated_at " + "FROM registration_projections WHERE entity_id = %s", + (node_id,), + ) + else: + cur.execute( + "SELECT entity_id, current_state, node_type, updated_at " + "FROM registration_projections ORDER BY updated_at DESC LIMIT 20" + ) + rows = cur.fetchall() except Exception as e:The same pattern applies to
_count_postgres_registrations(lines 437-450).
220-304: Hardcoded 5-second sleep may cause flaky tests.Line 285 uses a fixed
time.sleep(5)to wait for registration processing. This could be too short in slow CI environments or wastefully long in fast environments. Consider making this configurable or implementing a polling approach with exponential backoff.tests/unit/cli/infra_test/test_verify_topics.py (1)
17-18: Missing@pytest.mark.unitmarker.Per coding guidelines, test files should use pytest markers for classification. Add the unit marker to the test class.
🏷️ Proposed fix to add pytest marker
+@pytest.mark.unit class TestOnexTopicPattern: """Test ONEX 5-segment topic naming regex."""As per coding guidelines: "Use pytest markers:
@pytest.mark.unit,@pytest.mark.integration,@pytest.mark.slow,@pytest.mark.chaos,@pytest.mark.serialfor test classification"tests/unit/cli/infra_test/test_introspect.py (1)
41-44: Weak assertion on timestamp—only checks truthiness.The test asserts
payload["timestamp"]is truthy but doesn't verify it's a valid ISO 8601 string. Consider validating the format to catch malformed timestamps.💡 Optional: Validate ISO 8601 format
+ from datetime import datetime + def test_has_timestamp(self) -> None: """Payload includes a timestamp.""" payload = _build_introspection_payload() - assert payload["timestamp"] + ts = payload["timestamp"] + assert ts + # Validate ISO 8601 format + datetime.fromisoformat(ts.replace("Z", "+00:00"))src/omnibase_infra/cli/infra_test/run_suite.py (1)
196-218: Consider extracting duplicated environment verification logic.The environment verification check (docker compose ps) is repeated nearly identically in all three suite functions. A shared helper would reduce duplication and ensure consistent behavior.
♻️ Suggested helper extraction
def _verify_environment_running( compose_file: str, project_name: str, require_runtime: bool = False, ) -> None: """Verify E2E environment is running, exit if not.""" result = subprocess.run( [ "docker", "compose", "-f", compose_file, "-p", project_name, "ps", "--status" if not require_runtime else "--services", "running" if not require_runtime else None, ], capture_output=True, text=True, check=False, ) # Handle validation logic...Also applies to: 276-295, 358-380
tests/unit/cli/infra_test/test_cli_scaffold.py (1)
16-122: Missing@pytest.mark.unitdecorator on test class.Consistent with the earlier review comment on
test_introspect.py, this test class should include the@pytest.mark.unitmarker unless auto-applied via conftest.As per coding guidelines: "Use pytest markers:
@pytest.mark.unit,@pytest.mark.integration,@pytest.mark.slow,@pytest.mark.chaos,@pytest.mark.serialfor test classification".
Extract duplicated _get_broker(), _get_consul_addr(), and _get_postgres_dsn() into a shared _helpers.py module. Add str() casts and isinstance() guards in test_introspect.py to satisfy mypy with the dict[str, object] return type. Remove redundant os.getenv broker resolution in introspect command (Click already handles the envvar).
…yload
The version parameter was accepted but ignored — the function always
hardcoded node_version to {1, 0, 0}. Now parses the version string
into major/minor/patch components.
Review iteration: 1/10
- Fix connection leaks: use try/finally with conn.close() for all psycopg2 connections in run_suite.py and verify.py - Add subprocess timeout to all subprocess.run calls to prevent CI hangs - Extract duplicated env verification into _verify_env_running() helper - Fix comment/code mismatch (30s comment vs 10s sleep) in failure suite - Centralize broker fallback via get_broker() in introspect.py - Add @pytest.mark.unit to all 3 test classes - Add test_custom_version to verify version parameter affects output - Strengthen timestamp assertion to validate ISO 8601 format
- Fixed: introspect.py:48 - Add try/except for version parsing with click.BadParameter on non-numeric input - Fixed: _helpers.py:50 - URL-encode user/password in get_postgres_dsn() via urllib.parse.quote_plus - Added: test_helpers.py - Unit tests for get_broker, get_consul_addr, get_postgres_dsn (including special character encoding) - Added: test_introspect.py - Test for invalid version rejection Review iteration: 1/10
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Fix all issues with AI agents
In `@src/omnibase_infra/cli/infra_test/_helpers.py`:
- Around line 11-12: The DSN construction in this module builds a Postgres URI
using raw username/password which fails for special characters; import
urllib.parse and URL-encode the credentials with urllib.parse.quote_plus when
constructing the DSN (apply to the DSN builder used around lines 35-50). Find
the function or variable that assembles the Postgres URI (e.g.,
get_postgres_dsn, build_dsn, or the DSN variable) and replace raw username and
password insertion with quote_plus(username) and quote_plus(password), ensuring
the final URI uses the encoded values.
In `@src/omnibase_infra/cli/infra_test/introspect.py`:
- Around line 47-50: The version parsing currently does int(parts[i]) directly
which raises ValueError for non-numeric inputs; update the parsing of
version/parts to validate each part (e.g., str.isdigit() or try/except around
int conversion) and handle conversion failures by falling back to the default
for that segment (major=1, minor=0, patch=0) and emitting a friendly CLI error
(raise SystemExit or use the CLI framework's bad-parameter error) when the
entire --version is invalid; apply the same guarded logic to the other
occurrence referenced at lines 106-107 (the variables: version, parts, major,
minor, patch).
In `@src/omnibase_infra/cli/infra_test/verify.py`:
- Around line 292-413: The checks in _verify_consul_registry and
_verify_postgres_registry currently treat an overall empty result as non-failure
but do not fail when a specific node_id is requested but no matching entries
exist; update both functions to explicitly fail when node_id is provided and no
entries match the filter. In _verify_consul_registry, after building
onex_services or keys and applying the node_id filter (the loop that uses "if
node_id and node_id not in ...: continue"), detect if no items were
added/matched and console.print a red error mentioning the missing node_id and
return False; also adjust the printed table/count to reflect only matched items.
In _verify_postgres_registry, after fetching rows from registration_projections,
filter rows by node_id (or use the node-specific query already present) and if
node_id is provided but the filtered rows list is empty, console.print a red
error and return False; update the final count to use the number of matched rows
when printing. Ensure you reference the functions _verify_consul_registry and
_verify_postgres_registry and the variables onex_services, keys, and rows when
making these changes.
…dation - Fixed: verify.py - verify registry --node-id returns FAIL when specified node not found (Consul KV, services, and PostgreSQL paths) - Fixed: _helpers.py - Validate POSTGRES_HOST has no '@', POSTGRES_PORT is numeric, and CONSUL_SCHEME is http/https - Fixed: run_suite.py - Narrow except Exception to psycopg2.Error in _wait_for_registration polling loop - Added: test_helpers.py - Tests for invalid scheme, host, and port Review iteration: 2/10
- Fixed: verify.py - Replace broad except Exception with httpx.HTTPError for Consul calls and psycopg2.Error for PostgreSQL calls - Fixed: run_suite.py - Move psycopg2 imports before try blocks in idempotency and failure suites; narrow except to psycopg2.Error - Fixed: _helpers.py - URL-encode POSTGRES_DATABASE in DSN via quote_plus for consistency with user/password encoding Review iteration: 3/10
…nses - Fixed: verify.py - Widen Consul exception handlers to catch (httpx.HTTPError, ValueError) so json.JSONDecodeError (a ValueError subclass) from non-JSON 2xx responses is handled gracefully Review iteration: 4/10
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Fix all issues with AI agents
In `@src/omnibase_infra/cli/infra_test/run_suite.py`:
- Around line 305-313: The idempotency check currently allows count==0 to pass;
update the Idempotency suite in run_suite.py to first poll for at least one
record (reuse _wait_for_registration(dsn, node_id, timeout=15) or equivalent)
and fail early if no registration appears, then fetch the count and assert there
is exactly one record (or assert count > 0 and count <= 1) instead of allowing
zero; ensure you raise SystemExit(1) with a clear message when the wait times
out so the suite doesn't falsely pass when no events were processed.
🧹 Nitpick comments (1)
src/omnibase_infra/cli/infra_test/verify.py (1)
376-377: Minor: Count message shows total keys instead of matched count.When filtering by
--node-id, line 376 displayslen(keys)(total keys) rather thanmatched(filtered count). This could be confusing since only matched keys appear in the table.💡 Suggested fix
- console.print(f" [green]{len(keys)} service key(s) found.[/green]") + display_count = matched if node_id else len(keys) + console.print(f" [green]{display_count} service key(s) found.[/green]")
Comment on lines
+305
to
+313
| console.print(f" Records found: {count}") | ||
|
|
||
| if count <= 1: | ||
| console.print("[bold green]Idempotency suite: PASS[/bold green]") | ||
| else: | ||
| console.print( | ||
| f"[bold red]Idempotency suite: FAIL (expected <= 1, found {count})[/bold red]" | ||
| ) | ||
| raise SystemExit(1) |
There was a problem hiding this comment.
Idempotency test may pass falsely when no events are processed.
The test asserts count <= 1, which passes when count == 0. If events weren't processed within the 5-second settle time (e.g., consumer lag, service not running), the test would pass without verifying idempotency. Consider polling until at least one record exists (similar to _wait_for_registration in the smoke suite), then asserting count == 1.
🛠️ Suggested approach
console.print(f" Records found: {count}")
- if count <= 1:
+ if count == 1:
console.print("[bold green]Idempotency suite: PASS[/bold green]")
+ elif count == 0:
+ console.print(
+ "[bold red]Idempotency suite: FAIL "
+ "(no records found - events may not have been processed)[/bold red]"
+ )
+ raise SystemExit(1)
else:
console.print(
f"[bold red]Idempotency suite: FAIL (expected <= 1, found {count})[/bold red]"
)
raise SystemExit(1)Alternatively, poll for at least one record before checking the count:
# Wait for at least one record to appear
if not _wait_for_registration(dsn, node_id, timeout=15):
console.print("[bold red]No registration found within timeout.[/bold red]")
raise SystemExit(1)
# Then check for duplicates
count = ...🤖 Prompt for AI Agents
In `@src/omnibase_infra/cli/infra_test/run_suite.py` around lines 305 - 313, The
idempotency check currently allows count==0 to pass; update the Idempotency
suite in run_suite.py to first poll for at least one record (reuse
_wait_for_registration(dsn, node_id, timeout=15) or equivalent) and fail early
if no registration appears, then fetch the count and assert there is exactly one
record (or assert count > 0 and count <= 1) instead of allowing zero; ensure you
raise SystemExit(1) with a clear message when the wait times out so the suite
doesn't falsely pass when no events were processed.
…checks - Convert all psycopg2.connect() to context managers in run_suite.py (3 sites) and verify.py (2 sites) to prevent connection leaks on exception - Add version input guards in introspect.py: reject empty strings, >3 segments, and negative numbers - Strengthen timestamp display in verify.py with type/range validation instead of bare truthiness check - Add 5 edge-case tests for version parsing (partial, extra segments, negative, empty)
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Fix all issues with AI agents
In `@src/omnibase_infra/cli/infra_test/run_suite.py`:
- Around line 271-277: The idempotency test is generating new payloads each
iteration so events differ; instead construct a single introspection payload
once and reuse it for all publishes. Update the loop in run_suite.py to build
the payload before the for i in range(repetitions) loop and call
_publish_introspection(broker, topic, node_id, payload) (or modify
_publish_introspection to accept an optional payload) so
correlation_id/timestamp remain identical across all iterations; keep the
existing failure handling and console prints unchanged.
In `@src/omnibase_infra/cli/infra_test/verify.py`:
- Around line 268-285: The idempotency check currently treats count == 0 as a
pass; update the logic in verify.py where get_postgres_dsn() and
_count_postgres_registrations(dsn, node_id) are used so that count == None still
raises an error, count == 1 is the only success case, and both count == 0 and
count > 1 are treated as failures: for count == 0 print a clear "[bold red]No
registration records observed (possible processing failure or consumer
lag).[/bold red]" message and exit with SystemExit(1), and for count > 1 keep
the existing duplicate-record failure path.
…sitives - Fix psycopg2 connection leaks with explicit conn.close() in try/finally - Ensure idempotency suite publishes byte-identical payloads - Fail when count=0 in idempotency checks (no false positives) - Show matched vs total count when filtering Consul KV by node-id
…ous-truth guard - Fixed: run_suite.py:308,469 - Wrap psycopg2.connect() inside try/except so connection failures produce clean CLI errors instead of raw tracebacks - Fixed: _helpers.py:31 - Add port.isdigit() guard to get_consul_addr(), matching the existing validation in get_postgres_dsn() - Fixed: env.py:108 - Guard _wait_for_healthy against vacuous-truth when no service lines contain tab separators (same bug class as e58a17b) Review iteration: 1/10
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Fix all issues with AI agents
In `@src/omnibase_infra/cli/infra_test/verify.py`:
- Around line 433-455: The closing Rich markup tag for state styling is
incorrectly built using state_style.strip('[') which yields "green]" and creates
invalid markup; update the construction in the table.add_row call that builds
the state cell (the variables state_style and state are used there) so you strip
both brackets (e.g., strip('[]')) or otherwise extract the bare style name from
state_style and then concatenate it into a proper closing tag like
f"[/{style_name}]" so the final string is a valid closing tag for the opening
state_style.
Comment on lines
+433
to
+455
| for row in rows: | ||
| entity_id, state, ntype, updated = row | ||
| state_style = "[green]" if state == "ACTIVE" else "[yellow]" | ||
|
|
||
| # Validate timestamp with type and range checks, not just truthiness | ||
| if isinstance(updated, datetime): | ||
| now = datetime.now(tz=UTC) | ||
| # Make tz-aware for comparison (assume UTC if naive) | ||
| ts = updated if updated.tzinfo else updated.replace(tzinfo=UTC) | ||
| if ts > now: | ||
| updated_display = f"[red]{updated} (future timestamp)[/red]" | ||
| else: | ||
| updated_display = str(updated) | ||
| elif updated is not None: | ||
| updated_display = f"[yellow]{updated} (not a datetime)[/yellow]" | ||
| else: | ||
| updated_display = "[yellow]missing[/yellow]" | ||
|
|
||
| table.add_row( | ||
| str(entity_id), | ||
| f"{state_style}{state}[/{state_style.strip('[')}", | ||
| str(ntype or ""), | ||
| updated_display, |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
head -500 src/omnibase_infra/cli/infra_test/verify.py | tail -100Repository: OmniNode-ai/omnibase_infra
Length of output: 3730
🏁 Script executed:
sed -n '430,460p' src/omnibase_infra/cli/infra_test/verify.pyRepository: OmniNode-ai/omnibase_infra
Length of output: 1220
Fix Rich markup closing tag for state styling.
state_style.strip('[') on "[green]" produces "green]", creating an invalid closing tag [/green]]. Use strip('[]') to correctly produce "green", then add the closing ] explicitly to generate valid markup.
🛠️ Proposed fix
- f"{state_style}{state}[/{state_style.strip('[')}",
+ f"{state_style}{state}[/{state_style.strip('[]')}]",🤖 Prompt for AI Agents
In `@src/omnibase_infra/cli/infra_test/verify.py` around lines 433 - 455, The
closing Rich markup tag for state styling is incorrectly built using
state_style.strip('[') which yields "green]" and creates invalid markup; update
the construction in the table.add_row call that builds the state cell (the
variables state_style and state are used there) so you strip both brackets
(e.g., strip('[]')) or otherwise extract the bare style name from state_style
and then concatenate it into a proper closing tag like f"[/{style_name}]" so the
final string is a valid closing tag for the opening state_style.
jonahgabriel
deleted the
jonah/omn-1934-phase-5-integration-test-cli-p2
branch
February 8, 2026 18:53
This was referenced Feb 10, 2026
3 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
onex-infra-testCLI for end-to-end validation of the ONEX registration pipeline using Click + Richintrospectcommand to publish test events via rpk, andverifycommands for Consul, PostgreSQL, and Kafka state validationTest plan
poetry run pytest tests/unit/cli/infra_test/onex-infra-test --helpshows all commandsonex-infra-test env upstarts docker-compose.e2e.yml and waits for healthonex-infra-test verify registrychecks Consul + PostgreSQLonex-infra-test verify topicsvalidates ONEX naming complianceonex-infra-test run --suite smokepasses with live infrastructureSummary by CodeRabbit
New Features
Tests
Chores