Add a random jitter to target state polls, and a config var to ignore update notifications and not poll immediately after startup

This commit does two related things:

* We make the poll interval a random time between 0.5 and 1.5 times the configured interval.
* We introduce the BALENA_SUPERVISOR_INSTANT_UPDATE_TRIGGER configuration variable, that defaults to true. If this variable is set
to false, then calls to /v1/update are ignored, and on startup the supervisor waits for a poll interval before getting the target state.

This will help especially on cases where there's a large number of devices on a single network. By disabling instant updates and setting a large
poll interval, we can now achieve a sitation where not all devices apply an update at the same time, which can help avoid
overwhelming the network.

Change-type: minor
Signed-off-by: Pablo Carranza Velez <pablo@balena.io>

src/api-binder.ts

@@ -363,7 +363,7 @@ export class APIBinder {
 					'Trying to start poll without initializing API client',
 				);
 			}
-			this.pollTargetState();
+			this.pollTargetState(true);
 			return null;
 		});
 	}
@@ -568,19 +568,27 @@ export class APIBinder {
 			});
 	}
 
-	private async pollTargetState(): Promise<void> {
+	private async pollTargetState(isInitialCall: boolean = false): Promise<void> {
 		// TODO: Remove the checkInt here with the config changes
-		let pollInterval = await this.config.get('appUpdatePollInterval');
+		const { appUpdatePollInterval, instantUpdates } = await this.config.getMany(
+			['appUpdatePollInterval', 'instantUpdates'],
+		);
 
-		try {
-			await this.getAndSetTargetState(false);
-			this.targetStateFetchErrors = 0;
-		} catch (e) {
-			pollInterval = Math.min(
-				pollInterval,
-				15000 * 2 ** this.targetStateFetchErrors,
-			);
-			++this.targetStateFetchErrors;
+		// We add jitter to the poll interval so that it's between 0.5 and 1.5 times
+		// the configured interval
+		let pollInterval = (0.5 + Math.random()) * appUpdatePollInterval;
+
+		if (instantUpdates || !isInitialCall) {
+			try {
+				await this.getAndSetTargetState(false);
+				this.targetStateFetchErrors = 0;
+			} catch (e) {
+				pollInterval = Math.min(
+					appUpdatePollInterval,
+					15000 * 2 ** this.targetStateFetchErrors,
+				);
+				++this.targetStateFetchErrors;
+			}
 		}
 
 		await Bluebird.delay(pollInterval);
@@ -912,7 +920,15 @@ export class APIBinder {
 		router.post('/v1/update', (req, res) => {
 			apiBinder.eventTracker.track('Update notification');
 			if (apiBinder.readyForUpdates) {
-				apiBinder.getAndSetTargetState(req.body.force, true).catch(_.noop);
+				this.config.get('instantUpdates').then(instantUpdates => {
+					if (instantUpdates) {
+						apiBinder.getAndSetTargetState(req.body.force, true).catch(_.noop);
+					} else {
+						console.log(
+							'Ignoring update notification because instant updates are disabled',
+						);
+					}
+				});
 			}
 			res.sendStatus(204);
 		});

src/config/schema-type.ts

@@ -66,6 +66,10 @@ export const schemaTypes = {
 		type: PermissiveNumber,
 		default: 60000,
 	},
+	instantUpdates: {
+		type: PermissiveBoolean,
+		default: true,
+	},
 	mixpanelToken: {
 		type: t.string,
 		default: constants.defaultMixpanelToken,

src/config/schema.ts

@@ -185,6 +185,11 @@ export const schema = {
 		mutable: true,
 		removeIfNull: false,
 	},
+	instantUpdates: {
+		source: 'db',
+		mutable: true,
+		removeIfNull: false,
+	},
 };
 
 export type Schema = typeof schema;

src/device-config.ts

@@ -68,6 +68,11 @@ export class DeviceConfig {
 			varType: 'int',
 			defaultValue: '60000',
 		},
+		instantUpdates: {
+			envVarName: 'SUPERVISOR_INSTANT_UPDATE_TRIGGER',
+			varType: 'bool',
+			defaultValue: 'true',
+		},
 		localMode: {
 			envVarName: 'SUPERVISOR_LOCAL_MODE',
 			varType: 'bool',

test/05-device-state.spec.coffee

@@ -21,6 +21,7 @@ mockedInitialConfig = {
 	'RESIN_SUPERVISOR_DELTA_RETRY_COUNT': '30'
 	'RESIN_SUPERVISOR_DELTA_RETRY_INTERVAL': '10000'
 	'RESIN_SUPERVISOR_DELTA_VERSION': '2'
+	'RESIN_SUPERVISOR_INSTANT_UPDATE_TRIGGER': 'true'
 	'RESIN_SUPERVISOR_LOCAL_MODE': 'false'
 	'RESIN_SUPERVISOR_LOG_CONTROL': 'true'
 	'RESIN_SUPERVISOR_OVERRIDE_LOCK': 'false'
@@ -40,6 +41,7 @@ testTarget1 = {
 			'SUPERVISOR_DELTA_RETRY_COUNT': '30'
 			'SUPERVISOR_DELTA_RETRY_INTERVAL': '10000'
 			'SUPERVISOR_DELTA_VERSION': '2'
+			'SUPERVISOR_INSTANT_UPDATE_TRIGGER': 'true'
 			'SUPERVISOR_LOCAL_MODE': 'false'
 			'SUPERVISOR_LOG_CONTROL': 'true'
 			'SUPERVISOR_OVERRIDE_LOCK': 'false'
@@ -122,6 +124,7 @@ testTargetWithDefaults2 = {
 			'SUPERVISOR_DELTA_RETRY_COUNT': '30'
 			'SUPERVISOR_DELTA_RETRY_INTERVAL': '10000'
 			'SUPERVISOR_DELTA_VERSION': '2'
+			'SUPERVISOR_INSTANT_UPDATE_TRIGGER': 'true'
 			'SUPERVISOR_LOCAL_MODE': 'false'
 			'SUPERVISOR_LOG_CONTROL': 'true'
 			'SUPERVISOR_OVERRIDE_LOCK': 'false'