Gateway Helm Values
3 minute read
These values are specific to the gateway chart (edgedelta/edgedelta-gateway). For values shared across all charts, see Shared Helm Values.
helm upgrade edgedelta-gateway edgedelta/edgedelta-gateway -i \
--set secretApiKey.value=<pipeline-id> \
-n edgedelta --create-namespace
For architecture details, see Kubernetes Manifests — Gateway Pipeline.
Deployment
The gateway is always a Deployment (no deployment.kind option — unlike the node agent chart).
| Variable | Default | Description |
|---|---|---|
deployment.replicas | 2 | Processor pod count. Mutually exclusive with autoscaling. |
Autoscaling
| Variable | Default | Description |
|---|---|---|
deployment.autoscaling.enabled | false | Create a HorizontalPodAutoscaler for gateway processor agents. |
deployment.autoscaling.external | false | true if using an external autoscaler like KEDA. |
deployment.autoscaling.minReplicas | 1 | Minimum replica count. |
deployment.autoscaling.maxReplicas | 10 | Maximum replica count. |
deployment.autoscaling.targetForCPUUtilizationPercentage | 85 | Target average CPU utilization. |
deployment.autoscaling.targetForMemoryUtilizationPercentage | (empty) | Target average memory utilization. |
deployment.autoscaling.customMetric | {} | Custom metrics for HPA. |
deployment.autoscaling.behavior.scaleUp.stabilizationWindowSeconds | 60 | Scale-up stabilization window. |
deployment.autoscaling.behavior.scaleUp.policies[0] | Percent/50/60s | Scale up by max 50% per minute. |
deployment.autoscaling.behavior.scaleDown.stabilizationWindowSeconds | 300 | Scale-down stabilization window. |
The gateway chart includes scale-up policies by default (unlike the node agent chart), providing more conservative scale-up behavior for gateway workloads.
Topology Spread Constraints
| Variable | Default | Description |
|---|---|---|
deployment.topologySpreadConstraints | [] | Spread processor pods across topology domains. |
Compactor Agent
| Variable | Default | Description |
|---|---|---|
compactorProps.enabled | true | Enable the compactor agent. |
compactorProps.port | 9199 | Listening port. |
compactorProps.replicas | 1 | Static replica count. |
compactorProps.goMemLimit | "" | Go runtime memory limit. |
compactorProps.resources | 2000m/2Gi limits, 200m/300Mi requests | Resource constraints. |
compactorProps.tolerations | {} | Node tolerations. |
compactorProps.nodeSelector | {} | Node selector. |
compactorProps.priorityClassName | "" | PriorityClass. |
compactorProps.autoscaling.* | (same as node chart) | HPA configuration. |
compactorProps.topologySpreadConstraints | [] | Topology spread constraints. |
Rollup Agent
| Variable | Default | Description |
|---|---|---|
rollUpProps.enabled | true | Enable the rollup agent. |
rollUpProps.port | 9200 | Listening port. |
rollUpProps.replicas | 2 | Static replica count. |
rollUpProps.goMemLimit | "" | Go runtime memory limit. |
rollUpProps.resources | 1000m/1Gi limits, 200m/256Mi requests | Resource constraints. |
rollUpProps.tolerations | {} | Node tolerations. |
rollUpProps.nodeSelector | {} | Node selector. |
rollUpProps.priorityClassName | "" | PriorityClass. |
rollUpProps.autoscaling.* | (same as node chart) | HPA configuration. |
rollUpProps.topologySpreadConstraints | [] | Topology spread constraints. |
Container Ports
Variable: ports (default: [])
Define container ports for incoming data.
ports:
- name: otlp-grpc
protocol: TCP
port: 4317
The gateway chart does not include
pushService. To expose a gateway pipeline externally, create a Kubernetes Service manually. See Service Exposure — Gateway for an example.
Example — manual Service for gateway:
apiVersion: v1
kind: Service
metadata:
name: edgedelta-gateway-otlp
namespace: edgedelta
spec:
type: ClusterIP
selector:
app.kubernetes.io/name: edgedelta
edgedelta/agent-type: processor
ports:
- name: otlp-grpc
port: 4317
targetPort: 4317
protocol: TCP
Replace the
selectorlabels to match your Helm release. Check withkubectl get pods -n edgedelta --show-labels.
Coordinator Connection
| Variable | Default | Description |
|---|---|---|
coordinatorProps.enabled | true | Connect gateway agents to a coordinator. |
coordinatorProps.endpoint | "" | Coordinator endpoint. Empty = auto-discover. |
Service Monitor
| Variable | Default | Description |
|---|---|---|
serviceMonitor.enabled | false | Create a ServiceMonitor for Prometheus. |
Values NOT in this chart
The following node agent values are not available in the gateway chart:
deployment.kind— gateway is always a DeploymentpushService.*— create a Service manually insteadingress.*tracerProps.*(eBPF)httpRecorderProps.*persistingCursorProps.*dockerContainerProps.*aggregatorProps.*targetAllocator.*forceReinstallApplications.*persistence.*— see Coordinator Values