feat: improve availability of HyperDX pods (#75)

This PR introduces several optional configurations aimed at improving the availability and resilience of HyperDX pods.

Additions:
• PodDisruptionBudget (disabled by default):
  Ensures a minimum number of replicas remain available during voluntary disruptions. Defaults to a minimum of 1 replica but is configurable.

• Optional priorityClassName support:
  Enables better control over pod scheduling and preemption by allowing assignment of a custom priority class.

• Optional resource requests and limits:
  Defines resources for the app container to:
  – Guarantee the pod receives the required CPU/memory.
  – Prevent resource-hungry containers from affecting node stability.

• Optional topologySpreadConstraints:
  Allows pods to be distributed across different nodes (using a specified topology key), reducing the impact of node-level failures.

All features are disabled or conservative by default to maintain backward compatibility and allow for gradual opt-in.

Author: jspranger

.changeset/nine-dingos-lead.md

@@ -0,0 +1,5 @@
+---
+"helm-charts": patch
+---
+
+feat: improve availability of HyperDX pods

charts/hdx-oss-v2/templates/hyperdx-deployment.yaml

@@ -36,6 +36,13 @@ spec:
       tolerations:
         {{- toYaml .Values.hyperdx.tolerations | nindent 8 }}
       {{- end }}
+      {{- if .Values.hyperdx.topologySpreadConstraints }}
+      topologySpreadConstraints:
+        {{- toYaml .Values.hyperdx.topologySpreadConstraints | nindent 8 }}
+      {{- end }}
+      {{- if .Values.hyperdx.priorityClassName }}
+      priorityClassName: {{ .Values.hyperdx.priorityClassName | quote }}
+      {{- end }}
       {{- if .Values.mongodb.enabled }}
       initContainers:
         - name: wait-for-mongodb
@@ -53,6 +60,10 @@ spec:
               containerPort: {{ .Values.hyperdx.apiPort }}
             - name: opamp-port
               containerPort: {{ .Values.hyperdx.opampPort }}
+          {{- if .Values.hyperdx.resources }}
+          resources:
+            {{- toYaml .Values.hyperdx.resources | nindent 12 }}
+          {{- end }}
           envFrom:
             - configMapRef:
                 name: {{ include "hdx-oss.fullname" . }}-app-config

charts/hdx-oss-v2/templates/hyperdx-pdb.yaml

@@ -0,0 +1,20 @@
+{{- if .Values.hyperdx.podDisruptionBudget.enabled }}
+apiVersion: policy/v1
+kind: PodDisruptionBudget
+metadata:
+  name: {{ include "hdx-oss.fullname" . }}-pdb
+  labels:
+    {{- include "hdx-oss.labels" . | nindent 4 }}
+    app: {{ include "hdx-oss.fullname" . }}
+    {{- if .Values.hyperdx.labels }}
+    {{- with .Values.hyperdx.labels }}
+    {{- toYaml . | nindent 4 }}
+    {{- end -}}
+    {{- end }}
+spec:
+  minAvailable: {{ .Values.hyperdx.podDisruptionBudget.minAvailable | default 1 }}
+  selector:
+    matchLabels:
+      {{- include "hdx-oss.selectorLabels" . | nindent 6 }}
+      app: {{ include "hdx-oss.fullname" . }}
+{{- end }}

charts/hdx-oss-v2/tests/app-deployment_test.yaml

@@ -47,4 +47,53 @@ tests:
     asserts:
       - equal:
           path: spec.replicas
-          value: 3 
+          value: 3
+
+  - it: should set topology spread constraints when specified
+    set:
+      hyperdx:
+        topologySpreadConstraints:
+          - maxSkew: 1
+            topologyKey: topology.kubernetes.io/zone
+            whenUnsatisfiable: DoNotSchedule
+            labelSelector:
+              matchLabels:
+                foo: bar
+    asserts:
+      - contains:
+          path: spec.template.spec.topologySpreadConstraints
+          content:
+            maxSkew: 1
+            topologyKey: topology.kubernetes.io/zone
+            whenUnsatisfiable: DoNotSchedule
+            labelSelector:
+              matchLabels:
+                foo: bar
+
+  - it: should set priority class name when specified
+    set:
+      hyperdx:
+        priorityClassName: system-node-critical
+    asserts:
+      - equal:
+          path: spec.template.spec.priorityClassName
+          value: system-node-critical
+
+  - it: should set container resources when specified
+    set:
+      hyperdx:
+        resources:
+          limits:
+            memory: 1Gi
+          requests:
+            cpu: 100m
+            memory: 256Mi
+    asserts:
+      - equal:
+          path: spec.template.spec.containers[0].resources
+          value:
+            limits:
+              memory: 1Gi
+            requests:
+              cpu: 100m
+              memory: 256Mi

charts/hdx-oss-v2/tests/app-pdb_test.yaml

@@ -0,0 +1,29 @@
+suite: Test HyperDX App PodDisruptionBudget
+templates:
+  - hyperdx-pdb.yaml
+tests:
+  - it: should render the app pdb when enabled
+    set:
+      hyperdx:
+        podDisruptionBudget:
+          enabled: true
+    asserts:
+      - isKind:
+          of: PodDisruptionBudget
+      - equal:
+          path: spec.minAvailable
+          value: 1
+      - matchRegex:
+          path: metadata.name
+          pattern: -pdb$
+
+  - it: should override minimum available replicas when specified
+    set:
+      hyperdx:
+        podDisruptionBudget:
+          enabled: true
+          minAvailable: 5
+    asserts:
+      - equal:
+          path: spec.minAvailable
+          value: 5

charts/hdx-oss-v2/values.yaml

@@ -191,6 +191,9 @@ hyperdx:
 
   replicas: 1
 
+  podDisruptionBudget:
+    enabled: false
+
   # Service configuration
   service:
     type: ClusterIP  # Use ClusterIP for security. For external access, use ingress with proper TLS and authentication