gitea_admin/sientia-dataops-laborious_temporal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #40 from Aignosi/fix/SIENTIAPDE-1811-fix

Browse Source 
SIENTIAPDE-1811: Update dependencies, gitignore, and strengthen OPC UA error handling
This commit is contained in:
vitor-aignosi
2026-05-19 14:01:37 -03:00
committed by GitHub
parent 7c1dae8ef6 cd1be2430a
commit fcc8920a8b
6 changed files with 266 additions and 61 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
.gitignore vendored
View File

@@ -54,3 +54,5 @@ catboost_info/
mlruns/
relatorio*
openspec/
.cursor/

30
docs/opc-communication.md
View File

@@ -12,7 +12,7 @@ Worker (long-lived)
├── connect / disconnect / validate_connection (read-only)
├── _connect_locked / _reconnect_locked (under _connection_lock)
├── write_data (single attempt per call)
└── background reconnect on Tier-1 Bad*
└── background reconnect on Tier-1 Bad*, protocol closed, or stale session
Temporal activity write_opc_data
└── OPC.manage_output_tags → write_data per tag (sequential per activity)
@@ -26,8 +26,8 @@ One worker process holds one `OpcRepository` instance per configured server. Mul
|-------|----------|
| Startup | `init_opc()` creates repositories and calls `connect()` → `_connect_locked()` |
| Steady state | `validate_connection()` is read-only (`protocol.state` only); `_session_ready` is checked in `write_data` |
| Tier-1 Bad* | `_start_reconnect_on_bad` → `_run_reconnect_on_bad` → `_reconnect_locked()` (respects `reconnection_interval`) |
| Write | `write_data()` checks `_session_ready`, validates, then one `get_node` + `write_value` |
| Tier-1 Bad* / protocol closed | `_start_reconnect` → `_run_reconnect` → `_reconnect_locked()` (respects `reconnection_interval`) |
| Write | `write_data()` checks reconnect task, `_session_ready`, validates protocol, then one `get_node` + `write_value` |
| Shutdown | `close()` disconnects all repositories |
### Session and channel timeouts
@@ -36,7 +36,7 @@ Requested session and secure-channel lifetime: **10 minutes** (`OPC_UA_SESSION_A
### Reconnection interval
`OPC_RECONNECTION_INTERVAL` is in **seconds** (default `120`). It gates **background** reconnect after Tier-1 `Bad*` (`last_reconnection_time` is updated only in `_reconnect_locked()`). It limits load on the OPC server when many workflows fail at once.
`OPC_RECONNECTION_INTERVAL` is in **seconds** (default `120`). It gates **background** reconnect after Tier-1 `Bad*`, closed protocol, or stale session (`last_reconnection_time` is updated only in `_reconnect_locked()`). It limits load on the OPC server when many workflows fail at once.
## Concurrency: connection lock and session readiness
@@ -61,15 +61,17 @@ Public `connect()` / `disconnect()` acquire the lock and call `_connect_locked()
**Write path (`write_data`):**
1. If `_session_ready` is cleared → **fail immediately** (`opc_error_kind=reconnect_in_progress`).
2. `validate_connection()` checks `protocol.state` only (read-only).
3. Single `get_node` + `write_value` (no retry).
1. If a reconnect task is **in flight** → **fail immediately** (`opc_error_kind=reconnect_in_progress`).
2. If `_session_ready` is cleared and no task is running → schedule reconnect (`SessionNotReady`); fail with `connection_lost` or `reconnect_in_progress` if a task started.
3. `validate_connection()` checks `protocol.state` only (read-only). If closed → schedule reconnect (`ProtocolClosed`) and fail with `opc_error_kind=connection_lost`.
4. Single `get_node` + `write_value` (no retry). Tier-1 `Bad*` on write also schedules reconnect.
**Reconnect path (`_run_reconnect_on_bad`):**
**Reconnect path (`_run_reconnect`):**
1. `_start_reconnect_on_bad` clears `_session_ready` and schedules the task when the interval allows.
1. `_start_reconnect` clears `_session_ready` and schedules the task when the interval allows and `_allow_reconnect` is true.
2. `async with _connection_lock:` → `_reconnect_locked()`.
3. `_session_ready` is set on successful `_open_session()`.
4. `disconnect()` sets `_allow_reconnect=False` so shutdown does not respawn sessions.
A second `_connect_locked()` while a session is already open raises `OpcSessionAlreadyConnectedError` (disconnect first).
@@ -77,9 +79,15 @@ A second `_connect_locked()` while a session is already open raises `OpcSessionA
**Future threads:** replace `asyncio.Lock` / `Event` with `threading` primitives or route all OPC I/O through one dedicated loop.
## Tier-1 `Bad*` errors and reconnect
## Reconnect triggers
When the server invalidates the session (e.g. `BadSessionIdInvalid`) but the client still sees transport as open, `write_data` fails once, records the OPC status in metrics, and **schedules** reconnect if:
Background reconnect is scheduled when:
- `validate_connection()` sees a closed or missing protocol (`ProtocolClosed`).
- `_session_ready` is clear after a failed reconnect (`SessionNotReady`).
- A write raises a Tier-1 `UaStatusCodeError` in `RECONNECTABLE_OPC_BAD_NAMES`.
For Tier-1 `Bad*` when the server invalidates the session (e.g. `BadSessionIdInvalid`) but the client still sees transport as open, `write_data` fails once, records the OPC status in metrics, and **schedules** reconnect if:
- The exception is a `UaStatusCodeError` whose name is in `RECONNECTABLE_OPC_BAD_NAMES` (see plan), and
- `reconnection_interval` has elapsed since `last_reconnection_time`, and

194
laborious/utils/repository/opc_repository.py
View File

@@ -195,8 +195,18 @@ class OpcRepository(SientiaMonitoring):
self._connection_lock = asyncio.Lock()
self._session_ready = asyncio.Event()
self._reconnect_task: asyncio.Task[None] | None = None
self._allow_reconnect = True
def _opc_debug_tags(self, session_id: str) -> dict[str, str]:
"""
Build Prometheus/log label tags for OPC session-scoped metrics.
Args:
session_id (str): OPC UA session token string.
Return:
dict[str, str]: Labels pod_id, server_name, runtime, opc_server_id, session_id.
"""
return {
'pod_id': str(getattr(self, 'pod_id', 'unknown')),
'server_name': self.server_name,
@@ -234,6 +244,12 @@ class OpcRepository(SientiaMonitoring):
).total_seconds() > self.reconnection_interval
def _not_connected_error(self) -> dict[str, Any]:
"""
Build the standard error payload when validate_connection finds no open protocol.
Return:
dict[str, Any]: Notification fields for OPC_CONNECTION_NOT_READY.
"""
return {
'notification_id': f'OPC_CONNECTION_NOT_READY_{self.id}',
'message': f'OPC server {self.id} is not connected',
@@ -491,8 +507,12 @@ class OpcRepository(SientiaMonitoring):
async def disconnect(self) -> None:
"""
Gracefully disconnect from the OPC server under the connection lock.
Disables background reconnect so late writes during worker shutdown do not
respawn sessions.
"""
async with self._connection_lock:
self._allow_reconnect = False
await self._disconnect_locked()
async def validate_connection(self) -> tuple[bool, dict[str, Any]]:
@@ -509,58 +529,77 @@ class OpcRepository(SientiaMonitoring):
self.error(f'OPC server {self.id} is not connected', self.metadata)
return False, self._not_connected_error()
async def _start_reconnect_on_bad(self, opc_status: str, session_id: str) -> None:
def _reconnect_task_in_progress(self) -> bool:
"""
Schedule a background reconnect if interval and task state allow it.
Return whether a background reconnect task is currently running.
Return:
bool: True when a reconnect task exists and has not finished.
"""
return self._reconnect_task is not None and not self._reconnect_task.done()
async def _start_reconnect(self, reason: str, session_id: str) -> None:
"""
Schedule a background reconnect when allowed by interval and task state.
Clears _session_ready before starting the task. No-op when _allow_reconnect is
False, the reconnection window has not elapsed, or a reconnect is already running.
Args:
opc_status (str): OPC UA status name that triggered reconnect.
reason (str): Trigger for reconnect (OPC status name or synthetic reason).
session_id (str): Session token before failure.
"""
if not self._allow_reconnect:
return
if not self._reconnection_window_elapsed():
self.warning(
f'OPC reconnect skipped reason=reconnection_window opc_server_id={self.id} '
f'opc_status={opc_status}',
f'reconnect_reason={reason}',
self.metadata,
)
return
if self._reconnect_task is not None and not self._reconnect_task.done():
if self._reconnect_task_in_progress():
self.warning(
f'OPC reconnect skipped reason=in_progress opc_server_id={self.id} '
f'opc_status={opc_status}',
f'reconnect_reason={reason}',
self.metadata,
)
return
self._session_ready.clear()
self.info(
f'OPC reconnect scheduled after opc_status={opc_status} opc_server_id={self.id} '
f'OPC reconnect scheduled reconnect_reason={reason} opc_server_id={self.id} '
f'old_session_id={session_id}',
self.metadata,
)
self._reconnect_task = asyncio.create_task(
self._run_reconnect_on_bad(opc_status, session_id)
)
self._reconnect_task = asyncio.create_task(self._run_reconnect(reason, session_id))
async def _run_reconnect_on_bad(self, opc_status: str, session_id: str) -> None:
async def _run_reconnect(self, reason: str, session_id: str) -> None:
"""
Tear down and re-establish the OPC UA session under the connection lock.
Background task that tears down and re-establishes the OPC UA session.
Args:
opc_status (str): OPC UA status that triggered reconnect.
session_id (str): Previous session token string.
reason (str): Trigger for reconnect (OPC status or ProtocolClosed).
session_id (str): Previous session token string for logging.
"""
try:
async with self._connection_lock:
self.info(
f'OPC reconnect started opc_status={opc_status} opc_server_id={self.id} '
f'OPC reconnect started reconnect_reason={reason} opc_server_id={self.id} '
f'old_session_id={session_id}',
self.metadata,
)
await self._reconnect_locked()
success, error = await self._reconnect_locked()
if not success:
self.error(
f'OPC reconnect failed reconnect_reason={reason} opc_server_id={self.id}',
self.metadata,
)
if error:
self.error(error.get('message', ''), self.metadata)
except Exception:
self.error(
f'OPC reconnect task failed opc_server_id={self.id} opc_status={opc_status}',
f'OPC reconnect task failed opc_server_id={self.id} reconnect_reason={reason}',
self.metadata,
)
self.error(traceback.format_exc(), self.metadata)
@@ -593,6 +632,14 @@ class OpcRepository(SientiaMonitoring):
async def _emit_opc_write_metric(
self, session_id: str, result: str, metadata: dict[str, Any] | None
) -> None:
"""
Emit opc_write_attempts_total for a single write attempt outcome.
Args:
session_id (str): OPC UA session token string, or "unknown".
result (str): Outcome label (OK, OPC status name, ProtocolClosed, etc.).
metadata (dict[str, Any] | None): Write context for model_id/model_name labels.
"""
await self.emit_metric(
metrics.OPC_WRITE_ATTEMPTS_TOTAL,
{
@@ -611,6 +658,20 @@ class OpcRepository(SientiaMonitoring):
opc_error_kind: str | None = None,
opc_status: str | None = None,
) -> dict[str, Any]:
"""
Build a structured error dict returned from failed write_data paths.
Args:
notification_id (str): Stable notification identifier.
message (str): Human-readable failure message.
level (NotificationLevel): Severity for downstream notifications.
attachment_content (str | None): Optional traceback or diagnostic text.
opc_error_kind (str | None): Classifier (session_bad, connection_lost, etc.).
opc_status (str | None): OPC UA status name or synthetic reason.
Return:
dict[str, Any]: Error payload consumed by the OPC activity layer.
"""
payload: dict[str, Any] = {
'notification_id': notification_id,
'message': message,
@@ -656,7 +717,7 @@ class OpcRepository(SientiaMonitoring):
f'model_name={metadata.get("model_name", "unknown")} node={node} phase={phase}',
metadata,
)
await self._start_reconnect_on_bad(opc_status, session_id)
await self._start_reconnect(opc_status, session_id)
return False, self._write_failure_payload(
notification_id=f'OPC_WRITE_DATA_ERROR_{self.id}',
message=f'Failed to {phase} on OPC server: {exc} | metadata: {metadata}',
@@ -665,32 +726,89 @@ class OpcRepository(SientiaMonitoring):
opc_status=opc_status,
)
async def _write_reconnect_in_progress(
self, metadata: dict[str, Any]
) -> tuple[bool, dict[str, Any]]:
"""
Fail a write because a background reconnect task is already running.
Args:
metadata (dict[str, Any]): Write context passed through to the activity.
Return:
tuple[bool, dict[str, Any]]: (False, error info with opc_error_kind reconnect_in_progress).
"""
await self._emit_opc_write_metric('unknown', 'ReconnectInProgress', metadata)
self.warning(
f'OPC write rejected reconnect_in_progress opc_server_id={self.id} '
f'model_id={metadata.get("model_id", "unknown")} '
f'model_name={metadata.get("model_name", "unknown")}',
metadata,
)
return False, {
'notification_id': f'OPC_WRITE_RECONNECT_IN_PROGRESS_{self.id}',
'message': f'OPC write skipped: reconnect in progress | metadata: {metadata}',
'block': 'opc_repository',
'level': NotificationLevel.WARNING,
'opc_error_kind': 'reconnect_in_progress',
}
async def _write_connection_lost(
self, metadata: dict[str, Any], opc_status: str
) -> tuple[bool, dict[str, Any]]:
"""
Fail a write after scheduling reconnect for a closed or stale session.
Args:
metadata (dict[str, Any]): Write context passed through to the activity.
opc_status (str): Synthetic reason (ProtocolClosed, SessionNotReady).
Return:
tuple[bool, dict[str, Any]]: (False, error info with opc_error_kind connection_lost).
"""
await self._emit_opc_write_metric('unknown', opc_status, metadata)
return False, self._write_failure_payload(
notification_id=f'OPC_WRITE_CONNECTION_LOST_{self.id}',
message=f'OPC write skipped: connection lost ({opc_status}) | metadata: {metadata}',
level=NotificationLevel.WARNING,
opc_error_kind='connection_lost',
opc_status=opc_status,
)
async def write_data(
self, node: str, value: Any, data_type: str, metadata: dict[str, Any]
) -> tuple[bool, dict[str, Any]]:
"""
Write data to OPC server with a single attempt and Tier-1 Bad* reconnect scheduling.
"""
if not self._session_ready.is_set():
await self._emit_opc_write_metric('unknown', 'ReconnectInProgress', metadata)
self.warning(
f'OPC write rejected reconnect_in_progress opc_server_id={self.id} '
f'model_id={metadata.get("model_id", "unknown")} '
f'model_name={metadata.get("model_name", "unknown")}',
metadata,
)
return False, {
'notification_id': f'OPC_WRITE_RECONNECT_IN_PROGRESS_{self.id}',
'message': f'OPC write skipped: reconnect in progress | metadata: {metadata}',
'block': 'opc_repository',
'level': NotificationLevel.WARNING,
'opc_error_kind': 'reconnect_in_progress',
}
Write data to OPC server with a single attempt and background reconnect scheduling.
is_connected, error = await self.validate_connection()
Reconnect is scheduled on Tier-1 Bad*, closed protocol, or stale session readiness.
There is no retry within the same call.
Args:
node (str): OPC UA node id to write.
value (Any): Value to convert and send.
data_type (str): Logical type key (float, int, bool, str, double).
metadata (dict[str, Any]): Activity context (model_id, model_name, etc.).
Return:
tuple[bool, dict[str, Any]]: (True, {response_time}) on success, or
(False, structured error info) on failure.
"""
if self._reconnect_task_in_progress():
return await self._write_reconnect_in_progress(metadata)
if not self._session_ready.is_set():
session_id = _opc_authentication_token_str(self.client)
await self._start_reconnect('SessionNotReady', session_id)
if self._reconnect_task_in_progress():
return await self._write_reconnect_in_progress(metadata)
return await self._write_connection_lost(metadata, 'SessionNotReady')
is_connected, _error = await self.validate_connection()
if not is_connected:
await self._emit_opc_write_metric('unknown', 'NotConnected', metadata)
return False, error
session_id = _opc_authentication_token_str(self.client)
await self._start_reconnect('ProtocolClosed', session_id)
return await self._write_connection_lost(metadata, 'ProtocolClosed')
start_time = time.time()
session_id = _opc_authentication_token_str(self.client)

1
laborious/worker/worker.py
View File

@@ -205,6 +205,7 @@ async def main():
activities.cleanup_minio_objects_expired,
activities.repeat_last_prediction,
activities.export_data_to_postgres,
activities.export_payload_to_postgres,
activities.write_metrics,
# Pi Web API
activities.write_pi_web_api_data,

2
requirements-light.txt
View File

@@ -3,7 +3,7 @@ psycopg2-binary
sqlalchemy
asyncua
redis
git+ssh://git@github.com/Aignosi/sientia-dataops-library.git@1.10.4
git+ssh://git@github.com/Aignosi/sientia-dataops-library.git@1.12.0
prometheus-client
botocore
boto3

96
tests/laborious/utils/repository/test_opc_repository.py
View File

@@ -257,6 +257,7 @@ async def test_disconnect(opc_repository, mock_client):
opc_repository._disconnection_fallback.assert_called_once()
assert opc_repository.client is None
assert opc_repository._allow_reconnect is False
@pytest.mark.asyncio
@@ -331,14 +332,19 @@ async def test_write_data_validate_connection_do_nothing(opc_repository):
@pytest.mark.asyncio
async def test_write_data_validate_connection_failed(opc_repository):
opc_repository.validate_connection = AsyncMock(return_value=(False, {}))
opc_repository.client = AsyncMock()
opc_repository.client = MagicMock()
opc_repository.client.uaclient.protocol = MagicMock(state='closed')
opc_repository._start_reconnect = AsyncMock()
result = await opc_repository.write_data('ns=2;s=TestNode', 42.0, 'float', metadata['metadata'])
is_success, error_data = await opc_repository.write_data(
'ns=2;s=TestNode', 42.0, 'float', metadata['metadata']
)
opc_repository.validate_connection.assert_called_once()
opc_repository.client.get_node.assert_not_called()
assert result == (False, {})
opc_repository._start_reconnect.assert_called_once()
assert opc_repository._start_reconnect.call_args.args[0] == 'ProtocolClosed'
assert is_success is False
assert error_data['opc_error_kind'] == 'connection_lost'
assert error_data['opc_status'] == 'ProtocolClosed'
@pytest.mark.asyncio
@@ -439,7 +445,7 @@ def test_is_reconnectable_opcua_bad():
async def test_write_data_bad_session_id_invalid_schedules_reconnect(opc_repository, mock_client):
opc_repository.validate_connection = AsyncMock(return_value=(True, {}))
opc_repository.client = mock_client
opc_repository._start_reconnect_on_bad = AsyncMock()
opc_repository._start_reconnect = AsyncMock()
mock_node = AsyncMock()
mock_client.get_node = MagicMock(return_value=mock_node)
mock_node.write_value.side_effect = BadSessionIdInvalid()
@@ -449,7 +455,7 @@ async def test_write_data_bad_session_id_invalid_schedules_reconnect(opc_reposit
)
mock_node.write_value.assert_called_once()
opc_repository._start_reconnect_on_bad.assert_called_once()
opc_repository._start_reconnect.assert_called_once()
assert is_success is False
assert error_data['opc_error_kind'] == 'session_bad'
assert error_data['opc_status'] == 'BadSessionIdInvalid'
@@ -476,15 +482,85 @@ async def test_write_data_reconnect_in_progress_immediate(opc_repository):
@pytest.mark.asyncio
async def test_start_reconnect_on_bad_skips_within_interval(opc_repository):
async def test_start_reconnect_skips_within_interval(opc_repository):
opc_repository.last_reconnection_time = datetime.now()
opc_repository.reconnection_interval = 3600
await opc_repository._start_reconnect_on_bad('BadSessionIdInvalid', 'tok')
await opc_repository._start_reconnect('BadSessionIdInvalid', 'tok')
assert opc_repository._reconnect_task is None
@pytest.mark.asyncio
async def test_write_data_protocol_closed_schedules_reconnect(opc_repository):
opc_repository.client = MagicMock()
opc_repository.client.uaclient.protocol = MagicMock(state='closed')
opc_repository._start_reconnect = AsyncMock()
is_success, error_data = await opc_repository.write_data(
'ns=2;s=TestNode', 42.0, 'float', metadata['metadata']
)
opc_repository._start_reconnect.assert_called_once()
assert opc_repository._start_reconnect.call_args.args[0] == 'ProtocolClosed'
assert is_success is False
assert error_data['opc_error_kind'] == 'connection_lost'
assert error_data['opc_status'] == 'ProtocolClosed'
@pytest.mark.asyncio
async def test_write_data_protocol_closed_skips_reconnect_within_interval(opc_repository):
opc_repository.client = MagicMock()
opc_repository.client.uaclient.protocol = MagicMock(state='closed')
opc_repository.last_reconnection_time = datetime.now()
opc_repository.reconnection_interval = 3600
is_success, error_data = await opc_repository.write_data(
'ns=2;s=TestNode', 42.0, 'float', metadata['metadata']
)
assert opc_repository._reconnect_task is None
assert is_success is False
assert error_data['opc_error_kind'] == 'connection_lost'
@pytest.mark.asyncio
async def test_write_data_after_failed_reconnect_schedules_again(opc_repository):
opc_repository._session_ready.clear()
opc_repository.reconnection_interval = 0
opc_repository.last_reconnection_time = None
opc_repository._reconnect_locked = AsyncMock(
return_value=(False, {'message': 'connect failed'})
)
await opc_repository.write_data('ns=2;s=TestNode', 42.0, 'float', metadata['metadata'])
await asyncio.sleep(0.1)
assert opc_repository._reconnect_locked.call_count == 1
assert not opc_repository._reconnect_task_in_progress()
await opc_repository.write_data('ns=2;s=TestNode', 42.0, 'float', metadata['metadata'])
await asyncio.sleep(0.1)
assert opc_repository._reconnect_locked.call_count == 2
@pytest.mark.asyncio
async def test_write_data_after_disconnect_does_not_schedule_reconnect(opc_repository, mock_client):
opc_repository.client = mock_client
proto = MagicMock()
proto.state = 'closed'
mock_client.uaclient = MagicMock(protocol=proto)
opc_repository._disconnection_fallback = AsyncMock(return_value=[])
await opc_repository.disconnect()
is_success, error_data = await opc_repository.write_data(
'ns=2;s=TestNode', 42.0, 'float', metadata['metadata']
)
assert opc_repository._reconnect_task is None
assert is_success is False
assert error_data['opc_error_kind'] == 'connection_lost'
@pytest.mark.asyncio
async def test_parallel_bad_writes_single_reconnect_task(opc_repository, mock_client):
opc_repository.validate_connection = AsyncMock(return_value=(True, {}))