fix(clusters): resolve ip by create ts

infra/default-builds/dockerfiles/test-ds-echo/src/main.rs

@@ -96,7 +96,9 @@ async fn echo_tcp_server(port: u16) -> Result<()> {
 
 	let listener = TcpListener::bind(&addr).await.context("bind failed")?;
 	loop {
-		let (socket, _) = listener.accept().await.context("accept failed")?;
+		let (socket, addr) = listener.accept().await.context("accept failed")?;
+		println!("connection: {addr}");
+
 		tokio::spawn(async move {
 			let mut framed = Framed::new(socket, BytesCodec::new());
 
@@ -106,6 +108,8 @@ async fn echo_tcp_server(port: u16) -> Result<()> {
 					.await
 					.expect("write failed");
 			}
+
+			println!("connection closed: {addr}");
 		});
 	}
 }

infra/default-builds/outputs/test-ds-echo-tag.txt

@@ -1 +1 @@
-test-ds-echo:1723071093
+test-ds-echo:1723150754

lib/chirp-workflow/core/src/ctx/workflow.rs

@@ -1004,41 +1004,43 @@ impl WorkflowCtx {
 		Ok(signal)
 	}
 
-	/// Checks if the given signal exists in the database.
-	pub async fn query_signal<T: Listen>(&mut self) -> GlobalResult<Option<T>> {
-		let event = self.relevant_history().nth(self.location_idx);
-
-		// Signal received before
-		let signal = if let Some(event) = event {
-			tracing::debug!(name=%self.name, id=%self.workflow_id, "replaying signal");
-
-			// Validate history is consistent
-			let Event::Signal(signal) = event else {
-				return Err(WorkflowError::HistoryDiverged(format!(
-					"expected {event} at {}, found signal",
-					self.loc(),
-				)))
-				.map_err(GlobalError::raw);
-			};
-
-			Some(T::parse(&signal.name, signal.body.clone()).map_err(GlobalError::raw)?)
-		}
-		// Listen for new message
-		else {
-			let ctx = ListenCtx::new(self);
-
-			match T::listen(&ctx).await {
-				Ok(res) => Some(res),
-				Err(err) if matches!(err, WorkflowError::NoSignalFound(_)) => None,
-				Err(err) => return Err(err).map_err(GlobalError::raw),
-			}
-		};
-
-		// Move to next event
-		self.location_idx += 1;
-
-		Ok(signal)
-	}
+	// TODO: Currently implemented wrong, if no signal is received it should still write a signal row to the
+	// database so that upon replay it again receives no signal
+	// /// Checks if the given signal exists in the database.
+	// pub async fn query_signal<T: Listen>(&mut self) -> GlobalResult<Option<T>> {
+	// 	let event = self.relevant_history().nth(self.location_idx);
+
+	// 	// Signal received before
+	// 	let signal = if let Some(event) = event {
+	// 		tracing::debug!(name=%self.name, id=%self.workflow_id, "replaying signal");
+
+	// 		// Validate history is consistent
+	// 		let Event::Signal(signal) = event else {
+	// 			return Err(WorkflowError::HistoryDiverged(format!(
+	// 				"expected {event} at {}, found signal",
+	// 				self.loc(),
+	// 			)))
+	// 			.map_err(GlobalError::raw);
+	// 		};
+
+	// 		Some(T::parse(&signal.name, signal.body.clone()).map_err(GlobalError::raw)?)
+	// 	}
+	// 	// Listen for new message
+	// 	else {
+	// 		let ctx = ListenCtx::new(self);
+
+	// 		match T::listen(&ctx).await {
+	// 			Ok(res) => Some(res),
+	// 			Err(err) if matches!(err, WorkflowError::NoSignalFound(_)) => None,
+	// 			Err(err) => return Err(err).map_err(GlobalError::raw),
+	// 		}
+	// 	};
+
+	// 	// Move to next event
+	// 	self.location_idx += 1;
+
+	// 	Ok(signal)
+	// }
 
 	pub async fn msg<M>(&mut self, tags: serde_json::Value, body: M) -> GlobalResult<()>
 	where

svc/pkg/cluster/src/ops/server/resolve_for_ip.rs

@@ -32,6 +32,9 @@ pub async fn cluster_server_resolve_for_ip(
 		WHERE
 			($1 OR cloud_destroy_ts IS NULL) AND
 			public_ip = ANY($2)
+		-- When more than one record is returned per IP, sort by create_ts so that the first record is the
+		-- latest server created
+		ORDER BY create_ts DESC
 		",
 		input.include_destroyed,
 		input.ips

svc/pkg/cluster/src/workflows/server/install/install_scripts/files/node_exporter.sh

@@ -31,7 +31,7 @@ User=node_exporter
 Group=node_exporter
 Type=simple
 # Reduce cardinality
-ExecStart=/usr/bin/node_exporter --collector.disable-defaults --collector.cpu --collector.netdev --collector.conntrack --collector.meminfo --collector.filesystem --collector.filesystem.mount-points-exclude=^/opt/nomad/
+ExecStart=/usr/bin/node_exporter --collector.disable-defaults --collector.cpu --collector.netdev --collector.conntrack --collector.meminfo --collector.filesystem --collector.filesystem.mount-points-exclude=^/opt/nomad/ --collector.netstat --collector.sockstat --collector.tcpstat --collector.network_route --collector.arp --collector.filefd --collector.interrupts --collector.softirqs --collector.processes
 Restart=always
 RestartSec=2