feat(gateway): OpcConnectionPool (per-server, reuse, idle-evict)

2026-07-14 09:35:53 -03:00
parent 09e42901d1
commit 5d30a5c547
2 changed files with 253 additions and 0 deletions
--- a/ingestor/connection/pool.py
+++ b/ingestor/connection/pool.py
@@ -0,0 +1,120 @@
+import asyncio
+import traceback
+
+from asyncua import Client
+from sientia_do.notifications.handlers import CoreNotificationHandler as NotificationHandler
+from sientia_do.observability.logger import Logger
+from sientia_do.observability.metrics_controller import MetricsController
+from sientia_do.observability.sientia_monitoring import SientiaMonitoring
+
+import ingestor.metrics as metrics
+from ingestor.connection.certs import CertProvider
+from ingestor.connection.clock import MonotonicClock
+from ingestor.connection.security import apply_security
+
+
+class OpcConnectionPool(SientiaMonitoring):
+    """Per-server OPC UA client cache (F0 capstone): reuses live connections,
+    evicts idle ones, and reconnects after a failed attempt (never caches it)."""
+
+    def __init__(
+        self,
+        cert_provider: CertProvider,
+        logger: Logger,
+        notification_handler: NotificationHandler,
+        metrics_controller: MetricsController,
+        metadata: dict | None = None,
+    ):
+        self.cert_provider = cert_provider
+        self.metadata = metadata or {}
+        self.clock = MonotonicClock()
+        self._connections: dict[str, tuple[Client, float]] = {}
+
+        SientiaMonitoring.__init__(
+            self,
+            logger=logger,
+            notification_handler=notification_handler,
+            metrics_controller=metrics_controller,
+        )
+
+    async def get(self, server_id: str, server_cfg: dict) -> Client:
+        """Return the cached client for `server_id`, connecting one if needed."""
+        cached = self._connections.get(server_id)
+        if cached is not None:
+            client, _ = cached
+            self._connections[server_id] = (client, self.clock.elapsed())
+            return client
+
+        try:
+            client = await self._connect(server_cfg)
+        except Exception:
+            await self.emit_metric(
+                metric_object=metrics.OPC_CONNECTIONS_FAILED,
+                tags={'pod_id': self.pod_id, 'server_name': server_id},
+            )
+            raise
+
+        self._connections[server_id] = (client, self.clock.elapsed())
+        await self._record_connected(server_id, server_cfg['url'])
+        return client
+
+    async def _connect(self, server_cfg: dict) -> Client:
+        client = Client(server_cfg['url'], timeout=10)
+        client.application_uri = server_cfg.get('server_uri')
+        if server_cfg.get('use_security'):
+            cert, key = await self.cert_provider.fetch()
+            await apply_security(client, server_cfg.get('server_uri'), cert, key, None)
+        await client.connect()
+        return client
+
+    async def _record_connected(self, server_id: str, url: str):
+        await self.emit_metric(
+            metric_object=metrics.OPC_CONNECTIONS_TOTAL,
+            tags={'pod_id': self.pod_id, 'server_name': server_id},
+        )
+        await self.emit_metric(
+            metric_object=metrics.OPC_CONNECTION_STATUS,
+            method='set',
+            value=1,
+            tags={'pod_id': self.pod_id, 'server_name': server_id, 'server_url': url},
+        )
+
+    async def release_idle(self, max_idle_s: float):
+        """Disconnect and drop clients unused for longer than `max_idle_s`."""
+        now = self.clock.elapsed()
+        idle_ids = [
+            sid for sid, (_, last_used) in self._connections.items() if now - last_used > max_idle_s
+        ]
+        for server_id in idle_ids:
+            await self._evict(server_id)
+
+    async def close_all(self):
+        """Disconnect every pooled client (e.g. on shutdown)."""
+        for server_id in list(self._connections):
+            await self._evict(server_id)
+
+    async def _evict(self, server_id: str):
+        client, _ = self._connections.pop(server_id)
+        await self._disconnect_with_retries(client, server_id)
+        await self.emit_metric(
+            metric_object=metrics.OPC_DISCONNECTIONS_TOTAL,
+            tags={'pod_id': self.pod_id, 'server_name': server_id},
+        )
+
+    async def _disconnect_with_retries(self, client: Client, server_id: str) -> list:
+        """Tries 5 times to disconnect, with a delay of 100ms x try (mirrors
+        OpcManager.disconnection_fallback)."""
+        error_stack = []
+        for i in range(5):
+            try:
+                await client.disconnect()
+                return []
+            except Exception as e:
+                self.logger.error(
+                    f'Failed to disconnect from {server_id}, attempt {i + 1} of 5: {e}'
+                )
+                error_stack.append(
+                    {'attempt': i + 1, 'error': str(e), 'traceback': traceback.format_exc()}
+                )
+                await asyncio.sleep(0.1 * i)
+        return error_stack