How to configure trident resource limits

[kdx99]

Describe the bug
We want to set resource limits (cpu, memory) to the trident deployment and daemonset (requests seem already be set by default in the operator). But the limits we configure in the trident orchestrator are not passed to the deployments and pods. It lloks as if they are ignored.

Environment
Provide accurate information about the environment to help us reproduce the issue.

• Trident version: 26.2.0
• Trident installation flags used:
• Container runtime: OpenShift/cri-o
• Kubernetes version: v1.32.10
• Kubernetes orchestrator: OpenShift 4.19.21
• Kubernetes enabled feature gates:
• OS: Red Hat Enterprise Linux CoreOS 9.6.20251203-0
• NetApp backend types: Ontap 9.16.1P11
• Other:

To Reproduce
Steps to reproduce the behavior:
oc get tridentorchestrators.trident.netapp.io trident -o jsonpath='{.resources}'
-> limits are seen
oc get deployment trident-controller -n trident -o jsonpath='{.spec.template.spec.containers[].resources}'
-> only requests are seen
oc edit deployment trident-controller -n trident
oc edit daemonset trident-node-linux -n trident
-> setting limits there
-> limits disappear again

Expected behavior
when we define resource limits in the tridentorchestrator definition we expect also limits in the resulting pods in the trident namespace (and if not so at least a documentation how we can get limits in there)
Additional context
Add any other context about the problem here.
oc get tridentorchestrators.trident.netapp.io trident -o jsonpath='{.status.currentInstallationParams.resources}'
{"controller":{"csi-attacher":{"requests":{"cpu":"2m","memory":"20Mi"}},"csi-provisioner":{"requests":{"cpu":"2m","memory":"20Mi"}},"csi-resizer":{"requests":{"cpu":"3m","memory":"20Mi"}},"csi-snapshotter":{"requests":{"cpu":"2m","memory":"20Mi"}},"trident-autosupport":{"requests":{"cpu":"1m","memory":"30Mi"}},"trident-main":{"requests":{"cpu":"10m","memory":"80Mi"}}},"node":{"linux":{"node-driver-registrar":{"requests":{"cpu":"1m","memory":"10Mi"}},"trident-main":{"requests":{"cpu":"10m","memory":"60Mi"}}},"windows":{"liveness-probe":{"requests":{"cpu":"2m","memory":"40Mi"}},"node-driver-registrar":{"requests":{"cpu":"6m","memory":"40Mi"}},"trident-main":{"requests":{"cpu":"10m","memory":"60Mi"}}}}}
-> no limits there although configured

[torirevilla]

Hi @kdx99, the Trident documentation does describe how to set limits in the same level as where the requests are set under each container. Please double check the trident orchestrator has the correct structure. If problems persist please add more details including the resource part of your TORC.
https://docs.netapp.com/us-en/trident/trident-get-started/kubernetes-customize-deploy.html#configuration-options

[kdx99]

Hi @torirevilla , yes I have seen the docu, that is our torc definition:

apiVersion: trident.netapp.io/v1
kind: TridentOrchestrator
metadata:
  annotations:
    argocd.argoproj.io/sync-options: ServerSideApply=true
    argocd.argoproj.io/sync-wave: "1"
  name: trident
resources:
  controller:
    csi-attacher:
      limits:
        cpu: 500m
        memory: 3Gi
      requests:
        cpu: 2m
        memory: 20Mi
    csi-provisioner:
      limits:
        cpu: 500m
        memory: 3Gi
      requests:
        cpu: 2m
        memory: 20Mi
    csi-resizer:
      limits:
        cpu: 500m
        memory: 3Gi
      requests:
        cpu: 3m
        memory: 20Mi
    csi-snapshotter:
      limits:
        cpu: 500m
        memory: 3Gi
      requests:
        cpu: 2m
        memory: 20Mi
    trident-autosupport:
      limits:
        cpu: 500m
        memory: 3Gi
      requests:
        cpu: 1m
        memory: 30Mi
    trident-main:
      limits:
        cpu: 500m
        memory: 3Gi
      requests:
        cpu: 10m
        memory: 80Mi
  node:
    linux:
      node-driver-registrar:
        limits:
          cpu: 500m
          memory: 3Gi
        requests:
          cpu: 1m
          memory: 10Mi
      trident-main:
        limits:
          cpu: 500m
          memory: 3Gi
        requests:
          cpu: 10m
          memory: 60Mi
    windows:
      liveness-probe:
        limits:
          cpu: 500m
          memory: 3Gi
        requests:
          cpu: 2m
          memory: 40Mi
      node-driver-registrar:
        limits:
          cpu: 500m
          memory: 3Gi
        requests:
          cpu: 6m
          memory: 40Mi
      trident-main:
        limits:
          cpu: 500m
          memory: 3Gi
        requests:
          cpu: 10m
          memory: 60Mi
spec:
  IPv6: false
  cloudProvider: ""
  debug: true
  imagePullPolicy: IfNotPresent
  imagePullSecrets: []
  imageRegistry: ""
  namespace: trident
  nodePrep: []
  silenceAutosupport: false
  windows: false

Nevertheless the pod have either no limit (trident-operator) or only resource requests (trident-node-linux):

$ oc get deployment trident-operator -o jsonpath='{.spec.template.spec.containers[].resources}{"\n"}'
{}
$ oc get daemonsets trident-node-linux -n trident -o jsonpath='{.spec.template.spec.containers[].resources}{"\n"}'
{"requests":{"cpu":"10m","memory":"60Mi"}}

[NA-Scott]

It took a couple of tries to get it right, but it works for me:

jumphost:trident:~$ oc version
Client Version: 4.18.20
Kustomize Version: v5.4.2
Server Version: 4.21.4
Kubernetes Version: v1.34.4

jumphost:trident:~$ oc get tridentversion
NAME VERSION
trident 26.02.0

jumphost:trident:~$ oc get torc trident -o jsonpath='{.status.currentInstallationParams.resources.node.linux.trident-main}{"\n"}'
{"limits":{"cpu":"15m","memory":"90Mi"},"requests":{"cpu":"10m","memory":"60Mi"}}

jumphost:trident:~$ oc get daemonsets trident-node-linux -n trident -o jsonpath='{.spec.template.spec.containers[].resources}{"\n"}'
{"limits":{"cpu":"15m","memory":"90Mi"},"requests":{"cpu":"10m","memory":"60Mi"}}

You have to make the changes in the spec section...

jumphost:trident:~$ oc get torc trident -o yaml
apiVersion: trident.netapp.io/v1
kind: TridentOrchestrator
metadata:
  creationTimestamp: "2026-04-03T13:22:13Z"
  generation: 4
  name: trident
  resourceVersion: "3879565"
  uid: bf71e438-320c-470e-8a1b-efdc106821fa
spec:
  IPv6: false
  debug: true
  enableConcurrency: true
  imagePullSecrets: []
  imageRegistry: ""
  namespace: trident
  nodePrep: []
  resources:
    node:
      linux:
        node-driver-registrar:
          limits:
            cpu: 2m
            memory: 20Mi
        trident-main:
          limits:
            cpu: 15m
            memory: 90Mi
  silenceAutosupport: false
status:
  acpVersion: v26.02.0
  currentInstallationParams:
    IPv6: "false"
    acpImage: ""
    autosupportHostname: ""
    autosupportImage: registry.connect.redhat.com/netapp/trident-autosupport@sha256:566667b5e321b6b3b067e704b8a84f6bfce0425866ccec69674e45a6d4c2b197
    autosupportInsecure: false
    autosupportProxy: ""
    autosupportSerialNumber: ""
    debug: "true"
    disableAuditLog: "true"
    enableACP: "false"
    enableConcurrency: "true"
    enableForceDetach: "false"
    hostNetwork: false
    httpRequestTimeout: 90s
    httpsMetrics: "false"
    imagePullPolicy: IfNotPresent
    imagePullSecrets: []
    imageRegistry: ""
    iscsiSelfHealingInterval: 5m0s
    iscsiSelfHealingWaitTime: 7m0s
    k8sTimeout: "180"
    kubeletDir: /var/lib/kubelet
    logFormat: text
    logLayers: ""
    logLevel: debug
    logWorkflows: ""
    nodePrep: null
    probePort: "17546"
    resources:
      controller:
        csi-attacher:
          requests:
            cpu: 2m
            memory: 20Mi
        csi-provisioner:
          requests:
            cpu: 2m
            memory: 20Mi
        csi-resizer:
          requests:
            cpu: 3m
            memory: 20Mi
        csi-snapshotter:
          requests:
            cpu: 2m
            memory: 20Mi
        trident-autosupport:
          requests:
            cpu: 1m
            memory: 30Mi
        trident-main:
          requests:
            cpu: 10m
            memory: 80Mi
      node:
        linux:
          node-driver-registrar:
            limits:
              cpu: 2m
              memory: 20Mi
            requests:
              cpu: 1m
              memory: 10Mi
          trident-main:
            limits:
              cpu: 15m
              memory: 90Mi
            requests:
              cpu: 10m
              memory: 60Mi
        windows:
          liveness-probe:
            requests:
              cpu: 2m
              memory: 40Mi
          node-driver-registrar:
            requests:
              cpu: 6m
              memory: 40Mi
          trident-main:
            requests:
              cpu: 10m
              memory: 60Mi
    silenceAutosupport: "false"
    tridentImage: registry.connect.redhat.com/netapp/trident@sha256:c40d71da1785cab0c5c340b8030e4fa166cede8737acc7eac3d2dfd5c0292dce
  message: Trident installed
  namespace: trident
  status: Installed
  version: v26.02.0

[kdx99]

Thanks a lot for the hint! It did the trick. I modified a bit for getting all pods limited:

(...)
  resources:
    controller:
      trident-main:
        limits:
          cpu: 500m
          memory: 3Gi
      trident-autosupport:
        limits:
          cpu: 500m
          memory: 3Gi
      csi-provisioner:
        limits:
          cpu: 500m
          memory: 3Gi
      csi-attacher:
        limits:
          cpu: 500m
          memory: 3Gi
      csi-resizer:
        limits:
          cpu: 500m
          memory: 3Gi
      csi-snapshotter:
        limits:
          cpu: 500m
          memory: 3Gi
    node:
      linux:
        node-driver-registrar:
          limits:
            cpu: 500m
            memory: 3Gi
        trident-main:
          limits:
            cpu: 500m
            memory: 3Gi

The only thing we don't get working yet are the trident operator pods:

oc get deployment trident-operator -n trident-operator -o jsonpath='{.spec.template.spec.containers[].resources}{"\n"}'
{}