feat(helm): support topology spread constraints

Author: nicktrn

apps/supervisor/src/env.ts

@@ -90,6 +90,7 @@ const Env = z.object({
   KUBERNETES_MEMORY_REQUEST_MIN_GB: z.coerce.number().min(0).default(0),
   KUBERNETES_MEMORY_REQUEST_RATIO: z.coerce.number().min(0).max(1).default(1), // Ratio of memory limit, so 1 = 100% of memory limit
   KUBERNETES_MEMORY_OVERHEAD_GB: z.coerce.number().min(0).optional(), // Optional memory overhead to add to the limit in GB
+  KUBERNETES_TOPOLOGY_SPREAD_CONSTRAINTS: z.string().optional(), // JSON string
 
   // Placement tags settings
   PLACEMENT_TAGS_ENABLED: BoolEnv.default(false),

apps/supervisor/src/workloadManager/kubernetes.ts

@@ -56,6 +56,22 @@ export class KubernetesWorkloadManager implements WorkloadManager {
     };
   }
 
+  private parseTopologySpreadConstraints(): k8s.V1TopologySpreadConstraint[] | null {
+    if (!env.KUBERNETES_TOPOLOGY_SPREAD_CONSTRAINTS) {
+      return null;
+    }
+
+    try {
+      return JSON.parse(env.KUBERNETES_TOPOLOGY_SPREAD_CONSTRAINTS);
+    } catch (error) {
+      this.logger.error("[KubernetesWorkloadManager] Failed to parse topology spread constraints", {
+        error: error instanceof Error ? error.message : String(error),
+        raw: env.KUBERNETES_TOPOLOGY_SPREAD_CONSTRAINTS,
+      });
+      return null;
+    }
+  }
+
   private stripImageDigest(imageRef: string): string {
     if (!env.KUBERNETES_STRIP_IMAGE_DIGEST) {
       return imageRef;
@@ -270,6 +286,8 @@ export class KubernetesWorkloadManager implements WorkloadManager {
   }
 
   get #defaultPodSpec(): Omit<k8s.V1PodSpec, "containers"> {
+    const topologySpreadConstraints = this.parseTopologySpreadConstraints();
+
     return {
       restartPolicy: "Never",
       automountServiceAccountToken: false,
@@ -281,6 +299,7 @@ export class KubernetesWorkloadManager implements WorkloadManager {
             },
           }
         : {}),
+      ...(topologySpreadConstraints ? { topologySpreadConstraints } : {}),
     };
   }
 

hosting/k8s/helm/Chart.yaml

@@ -2,7 +2,7 @@ apiVersion: v2
 name: trigger
 description: The official Trigger.dev Helm chart
 type: application
-version: 4.0.1
+version: 4.0.2
 appVersion: v4.0.4
 home: https://trigger.dev
 sources:

hosting/k8s/helm/templates/supervisor.yaml

@@ -184,6 +184,10 @@ spec:
               value: {{ default "10Gi" .Values.supervisor.config.kubernetes.ephemeralStorageSizeLimit | quote }}
             - name: KUBERNETES_EPHEMERAL_STORAGE_SIZE_REQUEST
               value: {{ default "2Gi" .Values.supervisor.config.kubernetes.ephemeralStorageSizeRequest | quote }}
+            {{- with .Values.supervisor.config.kubernetes.topologySpreadConstraints }}
+            - name: KUBERNETES_WORKER_TOPOLOGY_SPREAD_CONSTRAINTS
+              value: {{ tpl (toYaml .) $ | toJson | quote }}
+            {{- end }}
             # Pod cleaner configuration
             - name: POD_CLEANER_ENABLED
               value: {{ .Values.supervisor.config.podCleaner.enabled | quote }}
@@ -272,6 +276,10 @@ spec:
       tolerations:
         {{- toYaml . | nindent 8 }}
       {{- end }}
+      {{- with .Values.supervisor.topologySpreadConstraints }}
+      topologySpreadConstraints:
+        {{- tpl (toYaml .) $ | nindent 8 }}
+      {{- end }}
 ---
 apiVersion: v1
 kind: Service

hosting/k8s/helm/templates/webapp.yaml

@@ -419,6 +419,10 @@ spec:
       tolerations:
         {{- toYaml . | nindent 8 }}
       {{- end }}
+      {{- with .Values.webapp.topologySpreadConstraints }}
+      topologySpreadConstraints:
+        {{- tpl (toYaml .) $ | nindent 8 }}
+      {{- end }}
 ---
 apiVersion: v1
 kind: Service

hosting/k8s/helm/values.yaml

@@ -69,6 +69,8 @@ webapp:
   nodeSelector: {}
   tolerations: []
   affinity: {}
+  # Topology Spread Constraints for pod assignment spread across your cluster among failure-domains. Evaluated as a template
+  topologySpreadConstraints: []
 
   logLevel: "info"
   gracefulShutdownTimeout: 1000
@@ -263,6 +265,8 @@ supervisor:
       workerNodetypeLabel: "" # When set, runs will only be scheduled on nodes with "nodetype=<label>"
       ephemeralStorageSizeLimit: "" # Default: 10Gi
       ephemeralStorageSizeRequest: "" # Default: 2Gi´
+      # Topology Spread Constraints for worker pods created by the supervisor. Evaluated as a template
+      topologySpreadConstraints: []
     podCleaner:
       enabled: true
       batchSize: 100
@@ -353,6 +357,8 @@ supervisor:
   nodeSelector: {}
   tolerations: []
   affinity: {}
+  # Topology Spread Constraints for pod assignment spread across your cluster among failure-domains. Evaluated as a template
+  topologySpreadConstraints: []
 
 # PostgreSQL configuration
 # Subchart: https://github.com/bitnami/charts/tree/main/bitnami/postgresql