fix(deployment): use hostname for topologyKey instead of region, update resources settings and staging configs (#7005)

Author: dotansimha

deployment/services/app.ts

@@ -46,7 +46,7 @@ export function deployApp({
     'app',
     {
       image,
-      replicas: environment.isProduction ? 3 : 1,
+      replicas: environment.podsConfig.general.replicas,
       imagePullSecret: docker.secret,
       readinessProbe: '/api/health',
       livenessProbe: '/api/health',

deployment/services/commerce.ts

@@ -49,7 +49,7 @@ export function deployCommerce({
     {
       image,
       imagePullSecret: docker.secret,
-      replicas: environment.isProduction ? 3 : 1,
+      replicas: environment.podsConfig.general.replicas,
       readinessProbe: '/_readiness',
       livenessProbe: '/_health',
       startupProbe: '/_health',

deployment/services/emails.ts

@@ -59,7 +59,7 @@ export function deployEmails({
       startupProbe: '/_health',
       exposesMetrics: true,
       image,
-      replicas: environment.isProduction ? 3 : 1,
+      replicas: environment.podsConfig.general.replicas,
     },
     [redis.deployment, redis.service],
   )

deployment/services/environment.ts

@@ -23,6 +23,9 @@ export function prepareEnvironment(input: {
 
   const appDns = `app.${input.rootDns}`;
   const apiDns = `api.${input.rootDns}`;
+  const isProduction = env === 'production';
+  const isStaging = env === 'staging';
+  const isDev = env === 'dev';
 
   return {
     envVars: {
@@ -33,14 +36,50 @@ export function prepareEnvironment(input: {
       RELEASE: input.release,
     },
     envName: env,
-    isProduction: env === 'production',
-    isStaging: env === 'staging',
-    isDev: env === 'dev',
+    isProduction,
+    isStaging,
+    isDev,
     encryptionSecret,
     release: input.release,
     appDns,
     apiDns,
     rootDns: input.rootDns,
+    podsConfig: {
+      general: {
+        replicas: isProduction ? 3 : isStaging ? 2 : 1,
+      },
+      supertokens: {
+        replicas: isProduction ? 3 : 1,
+      },
+      envoy: {
+        replicas: isProduction ? 3 : 1,
+        cpuLimit: isProduction ? '800m' : '150m',
+        memoryLimit: isProduction ? '1Gi' : '200Mi',
+      },
+      schemaService: {
+        memoryLimit: isProduction ? '2Gi' : '1Gi',
+      },
+      usageService: {
+        replicas: isProduction ? 3 : isStaging ? 2 : 1,
+        cpuLimit: isProduction ? '900m' : '300m',
+        maxReplicas: isProduction ? 6 : isStaging ? 3 : 1,
+        cpuAverageToScale: 60,
+      },
+      usageIngestorService: {
+        replicas: isProduction ? 6 : isStaging ? 2 : 1,
+        cpuLimit: isProduction ? '900m' : '300m',
+        maxReplicas: isProduction ? /* numberOfPartitions */ 16 : 2,
+        cpuAverageToScale: 60,
+      },
+      redis: {
+        memoryLimit: isProduction ? '4Gi' : '100Mi',
+        cpuLimit: isProduction ? '1000m' : '50m',
+      },
+      internalObservability: {
+        cpuLimit: isProduction ? '512m' : '150m',
+        memoryLimit: isProduction ? '1000Mi' : '300Mi',
+      },
+    },
   };
 }
 

deployment/services/graphql.ts

@@ -107,7 +107,7 @@ export function deployGraphQL({
       {
         imagePullSecret: docker.secret,
         image,
-        replicas: environment.isProduction ? 3 : 1,
+        replicas: environment.podsConfig.general.replicas,
         pdb: true,
         readinessProbe: '/_readiness',
         livenessProbe: '/_health',

deployment/services/policy.ts

@@ -32,7 +32,7 @@ export function deploySchemaPolicy({
     livenessProbe: '/_health',
     startupProbe: '/_health',
     exposesMetrics: true,
-    replicas: environment.isProduction ? 3 : 1,
+    replicas: environment.podsConfig.general.replicas,
     pdb: true,
   })
     .withConditionalSecret(sentry.enabled, 'SENTRY_DSN', sentry.secret, 'dsn')

deployment/services/proxy.ts

@@ -32,9 +32,9 @@ export function deployProxy({
   })
     .deployProxy({
       envoy: {
-        replicas: environment.isProduction ? 3 : 1,
-        cpu: environment.isProduction ? '800m' : '150m',
-        memory: environment.isProduction ? '800Mi' : '192Mi',
+        replicas: environment.podsConfig.envoy.replicas,
+        cpu: environment.podsConfig.envoy.cpuLimit,
+        memory: environment.podsConfig.envoy.memoryLimit,
       },
       tracing: observability.enabled
         ? { collectorService: observability.observability!.otlpCollectorService }

deployment/services/public-graphql-api-gateway.ts

@@ -60,7 +60,7 @@ export function deployPublicGraphQLAPIGateway(args: {
     {
       imagePullSecret: args.docker.secret,
       image: dockerImage,
-      replicas: args.environment.isProduction ? 3 : 1,
+      replicas: args.environment.podsConfig.general.replicas,
       availabilityOnEveryNode: true,
       env: {
         GRAPHQL_SERVICE_ENDPOINT: serviceLocalEndpoint(args.graphql.service).apply(

deployment/services/redis.ts

@@ -19,15 +19,10 @@ export function deployRedis(input: { environment: Environment }) {
   const redisApi = new RedisStore({
     password: redisPassword,
   }).deploy({
-    limits: input.environment.isProduction
-      ? {
-          memory: '6Gi',
-          cpu: '1000m',
-        }
-      : {
-          memory: '100Mi',
-          cpu: '50m',
-        },
+    limits: {
+      memory: input.environment.podsConfig.redis.memoryLimit,
+      cpu: input.environment.podsConfig.redis.cpuLimit,
+    },
   });
 
   const host = serviceLocalHost(redisApi.service);

deployment/services/schema.ts

@@ -51,8 +51,8 @@ export function deploySchema({
       livenessProbe: '/_health',
       startupProbe: '/_health',
       exposesMetrics: true,
-      replicas: environment.isProduction ? 3 : 1,
-      memoryLimit: '2Gi',
+      replicas: environment.podsConfig.general.replicas,
+      memoryLimit: environment.podsConfig.schemaService.memoryLimit,
       pdb: true,
     },
     [redis.deployment, redis.service],