fix: Prevent AWS_ERROR_MQTT_UNEXPECTED_HANGUP connection interruptions (#179)

By using a stable, unique per-process client identifier.

Also:

- Configured MQTT auto-reconnect on interruption
- Reset connection on high MQTT connection interruption rate

.github/dependabot.yml

@@ -8,7 +8,7 @@ updates:
   - package-ecosystem: 'npm' # See documentation for possible values
     directory: '/' # Location of package manifests
     schedule:
-      interval: 'daily'
+      interval: 'weekly'
     labels:
       - 'dependencies'
     commit-message:

example/main.ts

@@ -65,9 +65,10 @@ async function main() {
     client.emitter.on('statusChanged', (status) => {
       try {
         const device = devices.DevicesObj[status.deviceId];
-        const watts = status.current !== undefined ? status.current * device.Voltage : undefined;
+        const watts =
+          status.current !== undefined && device.Voltage !== undefined ? status.current * device.Voltage : undefined;
         rootLogger.info(
-          `[${status.deviceId}] '${device.Name}' status changed: ${status.temperature}°C, ${status.humidity}%, ${watts ?? 'na'}W`
+          `[${status.deviceId}] '${device.Name ?? 'Unknown'}' status changed: ${status.temperature}°C, ${status.humidity}%, ${watts ?? 'na'}W`
         );
       } catch (error) {
         rootLogger.error(error, `Error processing status update for device '${status.deviceId}'`);
@@ -77,7 +78,9 @@ async function main() {
     client.emitter.on('setPointChanged', (change) => {
       try {
         const device = devices.DevicesObj[change.deviceId];
-        rootLogger.info(`'${device.Name}' setpoint changed from ${change.previousSetPoint} to ${change.newSetPoint}`);
+        rootLogger.info(
+          `'${device.Name ?? 'Unknown'}' setpoint changed from ${change.previousSetPoint} to ${change.newSetPoint}`
+        );
       } catch (error) {
         rootLogger.error(error, `Error processing setpoint update for device '${change.deviceId}'`);
       }
@@ -86,7 +89,7 @@ async function main() {
     client.emitter.on('stateChanged', (change) => {
       try {
         const device = devices.DevicesObj[change.deviceId];
-        rootLogger.info(change, `'${device.Name}' state changed.`);
+        rootLogger.info(change, `'${device.Name ?? 'Unknown'}' state changed.`);
       } catch (error) {
         rootLogger.error(error, `Error processing state update for device '${change.deviceId}'`);
       }
@@ -96,7 +99,7 @@ async function main() {
   await Promise.all(
     Object.entries(devices.DevicesObj).map(async ([deviceId, device]) => {
       const serial = await client.getDeviceSerialNumber(deviceId);
-      rootLogger.info(`Serial number for device '${deviceId}' (${device.Name}): ${serial}`);
+      rootLogger.info(`Serial number for device '${deviceId}' (${device.Name ?? 'Unknown'}): ${serial}`);
 
       await client.startRealtimeUpdates(deviceId);
     })

package.json

@@ -51,20 +51,21 @@
   },
   "dependencies": {
     "@aws-sdk/client-iot": "3.920.0",
-    "@aws-sdk/credential-providers": "3.901.0",
+    "@aws-sdk/credential-providers": "3.922.0",
     "amazon-cognito-identity-js": "6.3.15",
     "aws-iot-device-sdk-v2": "1.22.0",
     "dayjs": "1.11.18",
-    "lodash": "4.17.21"
+    "lodash": "4.17.21",
+    "nanoid": "5.1.6"
   },
   "devDependencies": {
     "@eslint/js": "9.38.0",
     "@semantic-release/npm": "13.1.1",
     "@types/lodash": "4.17.20",
-    "@types/node": "24.9.2",
+    "@types/node": "24.10.0",
     "conventional-changelog-conventionalcommits": "9.1.0",
     "dotenv": "17.2.3",
-    "eslint": "9.38.0",
+    "eslint": "9.39.0",
     "eslint-plugin-jsdoc": "61.1.11",
     "eslint-plugin-tsdoc": "0.4.0",
     "pino": "9.13.0",
@@ -75,9 +76,9 @@
     "semantic-release": "25.0.1",
     "tsup": "8.5.0",
     "tsx": "4.20.6",
-    "typedoc": "0.28.13",
-    "typedoc-material-theme": "1.4.0",
+    "typedoc": "0.28.14",
+    "typedoc-material-theme": "1.4.1",
     "typescript": "5.9.3",
-    "typescript-eslint": "8.41.0"
+    "typescript-eslint": "8.46.2"
   }
 }

src/api/MysaApiClient.ts

@@ -21,6 +21,7 @@ import {
 import { iot, mqtt } from 'aws-iot-device-sdk-v2';
 import dayjs from 'dayjs';
 import duration from 'dayjs/plugin/duration.js';
+import { customAlphabet } from 'nanoid';
 import { MqttPublishError, MysaApiError, UnauthenticatedError } from './Errors';
 import { Logger, VoidLogger } from './Logger';
 import { MysaApiClientEventTypes } from './MysaApiClientEventTypes';
@@ -29,6 +30,8 @@ import { MysaDeviceMode, MysaFanSpeedMode } from './MysaDeviceMode';
 
 dayjs.extend(duration);
 
+const getRandomClientId = customAlphabet('1234567890abcdefghijklmnopqrstuvwxyz', 8);
+
 /** Options for MQTT publish operations. */
 export interface MqttPublishOptions {
   /** Maximum number of publish attempts before failing (default: 5). */
@@ -86,6 +89,12 @@ export class MysaApiClient {
   /** A promise that resolves to the MQTT connection used for real-time updates. */
   private _mqttConnectionPromise?: Promise<mqtt.MqttClientConnection>;
 
+  /** Stable per-process MQTT client id (prevents collisions between multiple processes). */
+  private _mqttClientId?: string;
+
+  /** Interrupt timestamps for storm / collision detection. */
+  private _mqttInterrupts: number[] = [];
+
   /** The device IDs that are currently being updated in real-time, mapped to their respective timeouts. */
   private _realtimeDeviceIds: Map<string, NodeJS.Timeout> = new Map();
 
@@ -173,6 +182,9 @@ export class MysaApiClient {
   async login(emailAddress: string, password: string): Promise<void> {
     this._cognitoUser = undefined;
     this._cognitoUserSession = undefined;
+    this._mqttClientId = undefined;
+    this._mqttInterrupts = [];
+
     this.emitter.emit('sessionChanged', this.session);
 
     return new Promise((resolve, reject) => {
@@ -601,6 +613,8 @@ export class MysaApiClient {
     const transientMarkers = [
       'AWS_ERROR_MQTT_TIMEOUT',
       'AWS_ERROR_MQTT_NO_CONNECTION',
+      'AWS_ERROR_MQTT_UNEXPECTED_HANGUP',
+      'UNEXPECTED_HANGUP',
       'Time limit between request and response',
       'timeout'
     ];
@@ -691,17 +705,21 @@ export class MysaApiClient {
     });
     const credentials = await credentialsProvider();
 
-    // Stable client id + persistent session to retain QoS1 queue & subscriptions across reconnects.
-    const stableClientId = `mysa-js-sdk-${this.session?.username ?? ''}`;
+    // Per-process stable client id. Random suffix avoids collisions with other running processes.
+    if (!this._mqttClientId) {
+      this._mqttClientId = `mysa-js-sdk-${this.session?.username ?? 'anon'}-${getRandomClientId()}`;
+    }
 
     const builder = iot.AwsIotMqttConnectionConfigBuilder.new_with_websockets()
       .with_credentials(AwsRegion, credentials.accessKeyId, credentials.secretAccessKey, credentials.sessionToken)
       .with_endpoint(MqttEndpoint)
-      .with_client_id(stableClientId)
+      .with_client_id(this._mqttClientId)
       .with_clean_session(false)
       .with_keep_alive_seconds(30)
       .with_ping_timeout_ms(3000)
-      .with_protocol_operation_timeout_ms(60000);
+      .with_protocol_operation_timeout_ms(60000)
+      .with_reconnect_min_sec(1)
+      .with_reconnect_max_sec(30);
 
     const config = builder.build();
     const client = new mqtt.MqttClient();
@@ -719,8 +737,38 @@ export class MysaApiClient {
       this._logger.error('MQTT connection_failure', e);
     });
 
-    connection.on('interrupt', (e) => {
+    connection.on('interrupt', async (e) => {
       this._logger.warn('MQTT interrupt', e);
+
+      // Track recent interrupts
+      const now = Date.now();
+
+      // Keep only last 60s
+      this._mqttInterrupts = this._mqttInterrupts.filter((t) => now - t < 60000);
+      this._mqttInterrupts.push(now);
+
+      if (this._mqttInterrupts.length > 5) {
+        this._logger.warn(
+          `High interrupt rate (${this._mqttInterrupts.length}/60s). Possible clientId collision. Regenerating clientId and resetting connection...`
+        );
+
+        // Force new client id to escape collision; close current connection
+        this._mqttClientId = undefined;
+
+        // Clear interrupts
+        this._mqttInterrupts = [];
+
+        try {
+          await connection.disconnect();
+
+          if (this._mqttConnectionPromise) {
+            this._logger.warn('MQTT connection promise still defined after disconnect; expected it to be cleared.');
+            this._mqttConnectionPromise = undefined;
+          }
+        } catch (error) {
+          this._logger.error('Failed to disconnect MQTT connection', error);
+        }
+      }
     });
 
     connection.on('resume', async (returnCode, sessionPresent) => {

src/types/rest/Devices.ts

@@ -10,9 +10,9 @@ export interface BrandInfo {
   /** Unique identifier for the brand */
   Id: number;
   /** Remote control model number for the AC device */
-  remoteModelNumber: string;
+  remoteModelNumber?: string;
   /** Original Equipment Manufacturer brand name */
-  OEMBrand: string;
+  OEMBrand?: string;
 }
 
 /**
@@ -56,55 +56,55 @@ export interface ModeObj {
  */
 export interface DeviceBase {
   /** Button digital input configuration value */
-  ButtonDI: number;
+  ButtonDI?: number;
   /** Maximum current rating as a string value */
-  MaxCurrent: string;
+  MaxCurrent?: string;
   /** Device model identifier string */
   Model: string;
   /** Button average value configuration */
-  ButtonAVE: number;
+  ButtonAVE?: number;
   /** Operating voltage of the device */
-  Voltage: number;
+  Voltage?: number;
   /** Button polling interval configuration */
-  ButtonPolling: number;
+  ButtonPolling?: number;
   /** Minimum brightness level (0-100) */
-  MinBrightness: number;
+  MinBrightness?: number;
   /** User-assigned device name */
-  Name: string;
+  Name?: string;
   /** Button low power mode configuration */
-  ButtonLowPower: number;
+  ButtonLowPower?: number;
   /** Type of heater controlled by the device */
-  HeaterType: string;
+  HeaterType?: string;
   /** Button repeat delay configuration in milliseconds */
-  ButtonRepeatDelay: number;
+  ButtonRepeatDelay?: number;
   /** Button repeat start delay configuration in milliseconds */
-  ButtonRepeatStart: number;
+  ButtonRepeatStart?: number;
   /** Display animation style setting */
-  Animation: string;
+  Animation?: string;
   /** Maximum brightness level (0-100) */
-  MaxBrightness: number;
+  MaxBrightness?: number;
   /** Array of user IDs allowed to control this device */
-  AllowedUsers: string[];
+  AllowedUsers?: string[];
   /** Current button state indicator */
-  ButtonState: string;
+  ButtonState?: string;
   /** Home identifier that this device belongs to */
-  Home: string;
+  Home?: string;
   /** Button sensitivity threshold configuration */
-  ButtonThreshold: number;
+  ButtonThreshold?: number;
   /** Data format version used by the device */
-  Format: string;
+  Format?: string;
   /** Time zone setting for the device */
-  TimeZone: string;
+  TimeZone?: string;
   /** Unix timestamp of when device was last paired */
-  LastPaired: number;
+  LastPaired?: number;
   /** Minimum temperature setpoint allowed */
-  MinSetpoint: number;
+  MinSetpoint?: number;
   /** Current operating mode of the device */
-  Mode: ModeObj;
+  Mode?: ModeObj;
   /** User ID of the device owner */
-  Owner: string;
+  Owner?: string;
   /** Maximum temperature setpoint allowed */
-  MaxSetpoint: number;
+  MaxSetpoint?: number;
   /** Unique device identifier */
   Id: string;
   /** Optional zone assignment for the device */

src/types/rest/States.ts

@@ -13,43 +13,43 @@ export interface DeviceState {
   /** Overall timestamp for the device state */
   Timestamp: number;
   /** Time the device has been on */
-  OnTime: TimestampedValue<number>;
+  OnTime?: TimestampedValue<number>;
   /** Temperature set point */
-  SetPoint: TimestampedValue<number>;
+  SetPoint?: TimestampedValue<number>;
   /** Display brightness level */
-  Brightness: TimestampedValue<number>;
+  Brightness?: TimestampedValue<number>;
   /** Schedule mode setting */
-  ScheduleMode: TimestampedValue<number>;
+  ScheduleMode?: TimestampedValue<number>;
   /** Hold time setting */
-  HoldTime: TimestampedValue<number>;
+  HoldTime?: TimestampedValue<number>;
   /** Wi-Fi signal strength */
-  Rssi: TimestampedValue<number>;
+  Rssi?: TimestampedValue<number>;
   /** Thermostat mode */
-  TstatMode: TimestampedValue<number>;
+  TstatMode?: TimestampedValue<number>;
   /** Available heap memory */
-  FreeHeap: TimestampedValue<number>;
+  FreeHeap?: TimestampedValue<number>;
   /** Sensor temperature reading */
-  SensorTemp: TimestampedValue<number>;
+  SensorTemp?: TimestampedValue<number>;
   /** Current mode */
-  Mode: TimestampedValue<number>;
+  Mode?: TimestampedValue<number>;
   /** Voltage measurement */
-  Voltage: TimestampedValue<number>;
+  Voltage?: TimestampedValue<number>;
   /** Temperature corrected for calibration */
-  CorrectedTemp: TimestampedValue<number>;
+  CorrectedTemp?: TimestampedValue<number>;
   /** Duty cycle percentage */
-  Duty: TimestampedValue<number>;
+  Duty?: TimestampedValue<number>;
   /** Heat sink temperature */
-  HeatSink: TimestampedValue<number>;
+  HeatSink?: TimestampedValue<number>;
   /** Time the device has been off */
-  OffTime: TimestampedValue<number>;
+  OffTime?: TimestampedValue<number>;
   /** Connection status */
-  Connected: TimestampedValue<boolean>;
+  Connected?: TimestampedValue<boolean>;
   /** Current consumption */
-  Current: TimestampedValue<number>;
+  Current?: TimestampedValue<number>;
   /** Humidity reading */
-  Humidity: TimestampedValue<number>;
+  Humidity?: TimestampedValue<number>;
   /** Lock status */
-  Lock: TimestampedValue<number>;
+  Lock?: TimestampedValue<number>;
   /** Fan speed */
   FanSpeed?: TimestampedValue<number>;
 }