feat(topology): support variable fabric and accelerated tiers

Signed-off-by: Dmitry Shmulevich <dshmulevich@nvidia.com>

Author: dmitsh

.claude/CLAUDE.md

@@ -55,8 +55,8 @@ These structures propagate across every provider and engine. Changing them in a
 | Surface | Why it's load-bearing |
 |---|---|
 | `pkg/topology/` — `Graph`, the `Vertex` tree, and topology constants | Every provider returns it; every engine consumes it. A shape change ripples to all of them. |
-| Helm `provider.name` / `engine.name` / `topologyNodeLabels` | External contract for operators deploying Topograph. |
-| The variable label families `network.topology.nvidia.com/level-N` and `accelerated.topology.nvidia.com/level-N` | Consumed by downstream projects (KAI Scheduler, NVSentinel, Kueue); level 0 is closest to the node. |
+| Helm `provider.name` / `engine.name` | External contract for operators deploying Topograph. |
+| The variable label families `network.topology.nvidia.com/tier-N` and `accelerated.topology.nvidia.com/domain-N` | Consumed by downstream projects (KAI Scheduler, NVSentinel, Kueue); fabric tier 0 is closest to the node and accelerated domain 0 is narrowest. |
 
 ## 2. Setup and Installation
 
@@ -169,7 +169,7 @@ Engines are much rarer (four exist: slurm, k8s, nfd, slinky). Follow the same re
 
 ### Label and annotation reference
 
-Label keys written by the Kubernetes and Slinky engines are documented in `docs/reference/node-labels.md`. Do not invent new keys in provider or engine code — values flow through the canonical graph; keys are configured via Helm `topologyNodeLabels`.
+Label keys written by the Kubernetes engine are documented in `docs/reference/node-labels.md`. Do not invent new keys in provider code — values flow through the canonical graph. Optional custom keys are configured independently through the k8s engine's closest-first `fabricLabels` and narrowest-first `acceleratedLabels` arrays; when an array is provided, only explicitly listed entries in that family are labeled.
 
 ## 5. Pull Request Guidelines
 

AGENTS.md

@@ -55,8 +55,8 @@ These structures propagate across every provider and engine. Changing them in a
 | Surface | Why it's load-bearing |
 |---|---|
 | `pkg/topology/` — `Graph`, the `Vertex` tree, and topology constants | Every provider returns it; every engine consumes it. A shape change ripples to all of them. |
-| Helm `provider.name` / `engine.name` / `topologyNodeLabels` | External contract for operators deploying Topograph. |
-| The variable label families `network.topology.nvidia.com/level-N` and `accelerated.topology.nvidia.com/level-N` | Consumed by downstream projects (KAI Scheduler, NVSentinel, Kueue); level 0 is closest to the node. |
+| Helm `provider.name` / `engine.name` | External contract for operators deploying Topograph. |
+| The variable label families `network.topology.nvidia.com/tier-N` and `accelerated.topology.nvidia.com/domain-N` | Consumed by downstream projects (KAI Scheduler, NVSentinel, Kueue); fabric tier 0 is closest to the node and accelerated domain 0 is narrowest. |
 
 ## 2. Setup and Installation
 
@@ -169,7 +169,7 @@ Engines are much rarer (four exist: slurm, k8s, nfd, slinky). Follow the same re
 
 ### Label and annotation reference
 
-Label keys written by the Kubernetes and Slinky engines are documented in `docs/reference/node-labels.md`. Do not invent new keys in provider or engine code — values flow through the canonical graph; keys are configured via Helm `topologyNodeLabels`.
+Label keys written by the Kubernetes engine are documented in `docs/reference/node-labels.md`. Do not invent new keys in provider code — values flow through the canonical graph. Optional custom keys are configured independently through the k8s engine's closest-first `fabricLabels` and narrowest-first `acceleratedLabels` arrays; when an array is provided, only explicitly listed entries in that family are labeled.
 
 ## 5. Pull Request Guidelines
 

CHANGELOG.md

@@ -19,7 +19,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 
 ### Changed
 
-- **BREAKING:** Topology now uses variable, closest-first level families: `network.topology.nvidia.com/level-N` for fabric tiers and `accelerated.topology.nvidia.com/level-N` for accelerated tiers. The canonical `InstanceTopology` and graph conversion support arbitrary depths; the fixed `accelerator`, `leaf`, `spine`, and `core` label keys and their Helm overrides are replaced by `topologyNodeLabels.fabricPrefix` and `topologyNodeLabels.acceleratedPrefix`.
+- **BREAKING:** Topology now uses variable, closest-first fabric tiers and narrowest-first accelerated domains: `network.topology.nvidia.com/tier-N` and `accelerated.topology.nvidia.com/domain-N`. The canonical `InstanceTopology` and graph conversion support arbitrary depths. The fixed `accelerator`, `leaf`, `spine`, and `core` keys and process-wide Helm/CLI overrides are replaced by optional `fabricLabels` and `acceleratedLabels` arrays on the `k8s` engine. Each omitted array defaults every discovered entry in its family; when provided, only explicitly configured entries in that family are labeled.
 - Simulation model node names are now treated as hostnames; the model-backed test provider generates their instance IDs with an `i-` prefix.
 - The node-observer and node-data-broker are now rendered directly by the main Topograph Helm chart instead of local subcharts. Their existing `node-observer.*` and `node-data-broker.*` values paths are unchanged.
 - **BREAKING (Helm chart `0.5.0` → `0.6.0`):** the chart now ships a hardened default security context across the API server, node-observer, and node-data-broker: non-root (`runAsNonRoot`, UID/GID `65532`), `seccompProfile: RuntimeDefault`, `allowPrivilegeEscalation: false`, `readOnlyRootFilesystem: true`, and all capabilities dropped — satisfying the Kubernetes `restricted` Pod Security Standard out of the box. This changes the default runtime posture of every workload; operators who relied on root, a writable rootfs, or added capabilities must override the relevant keys (see the migration note below). `appVersion` is unchanged (`v0.5.0`; no binary change).

charts/topograph/templates/NOTES.txt

@@ -48,9 +48,6 @@
         {{ .Values.nfdNamespace }}. This must be the namespace where NFD master runs.
 {{- end }}
 
-  Topology label families use closest-first numeric levels:
-    fabric:      {{ .Values.topologyNodeLabels.fabricPrefix }}N
-    accelerated: {{ .Values.topologyNodeLabels.acceleratedPrefix }}N
 {{- if not .Values.serviceAccount.create }}
 
 2. Topograph is configured to use the existing ServiceAccount "{{ include "topograph.serviceAccountName" . }}".

charts/topograph/templates/deployment.yaml

@@ -50,12 +50,6 @@ spec:
             - /usr/local/bin/topograph
           args:
             - -v={{ .Values.verbosity }}
-            {{- with .Values.topologyNodeLabels.fabricPrefix }}
-            - -k8s-topology-key-fabric-prefix={{ . }}
-            {{- end }}
-            {{- with .Values.topologyNodeLabels.acceleratedPrefix }}
-            - -k8s-topology-key-accelerated-prefix={{ . }}
-            {{- end }}
           {{- if or .Values.nodeDataBroker.enabled $hasGcpCredentials $isNFDEngine .Values.env }}
           env:
             {{- if .Values.nodeDataBroker.enabled }}

charts/topograph/tests/__snapshot__/render_snapshot_test.yaml.snap

@@ -8,10 +8,6 @@ renders default values.yaml:
         kubectl --namespace topograph port-forward $POD_NAME 8080:$CONTAINER_PORT
 
         NOTE: node-data-broker applies node annotations once when each broker pod starts.
-
-        Topology label families use closest-first numeric levels:
-          fabric:      network.topology.nvidia.com/level-N
-          accelerated: accelerated.topology.nvidia.com/level-N
   2: |
     apiVersion: v1
     data:
@@ -63,8 +59,6 @@ renders default values.yaml:
           containers:
             - args:
                 - -v=3
-                - -k8s-topology-key-fabric-prefix=network.topology.nvidia.com/level-
-                - -k8s-topology-key-accelerated-prefix=accelerated.topology.nvidia.com/level-
               command:
                 - /usr/local/bin/topograph
               env:
@@ -557,10 +551,6 @@ renders values.k8s.gateway-api-example.yaml:
               it and (if needed) published its listener address in DNS.
 
         NOTE: node-data-broker applies node annotations once when each broker pod starts.
-
-        Topology label families use closest-first numeric levels:
-          fabric:      network.topology.nvidia.com/level-N
-          accelerated: accelerated.topology.nvidia.com/level-N
   2: |
     apiVersion: v1
     data:
@@ -612,8 +602,6 @@ renders values.k8s.gateway-api-example.yaml:
           containers:
             - args:
                 - -v=3
-                - -k8s-topology-key-fabric-prefix=network.topology.nvidia.com/level-
-                - -k8s-topology-key-accelerated-prefix=accelerated.topology.nvidia.com/level-
               command:
                 - /usr/local/bin/topograph
               env:
@@ -1128,10 +1116,6 @@ renders values.k8s.gcp-federated-workload-identity-example.yaml:
         kubectl --namespace topograph port-forward $POD_NAME 8080:$CONTAINER_PORT
 
         NOTE: node-data-broker applies node annotations once when each broker pod starts.
-
-        Topology label families use closest-first numeric levels:
-          fabric:      network.topology.nvidia.com/level-N
-          accelerated: accelerated.topology.nvidia.com/level-N
   2: |
     apiVersion: v1
     data:
@@ -1183,8 +1167,6 @@ renders values.k8s.gcp-federated-workload-identity-example.yaml:
           containers:
             - args:
                 - -v=3
-                - -k8s-topology-key-fabric-prefix=network.topology.nvidia.com/level-
-                - -k8s-topology-key-accelerated-prefix=accelerated.topology.nvidia.com/level-
               command:
                 - /usr/local/bin/topograph
               env:
@@ -1709,10 +1691,6 @@ renders values.k8s.gcp-service-account-example.yaml:
         kubectl --namespace topograph port-forward $POD_NAME 8080:$CONTAINER_PORT
 
         NOTE: node-data-broker applies node annotations once when each broker pod starts.
-
-        Topology label families use closest-first numeric levels:
-          fabric:      network.topology.nvidia.com/level-N
-          accelerated: accelerated.topology.nvidia.com/level-N
   2: |
     apiVersion: v1
     data:
@@ -1764,8 +1742,6 @@ renders values.k8s.gcp-service-account-example.yaml:
           containers:
             - args:
                 - -v=3
-                - -k8s-topology-key-fabric-prefix=network.topology.nvidia.com/level-
-                - -k8s-topology-key-accelerated-prefix=accelerated.topology.nvidia.com/level-
               command:
                 - /usr/local/bin/topograph
               env:
@@ -2278,10 +2254,6 @@ renders values.k8s.ib-example.yaml:
         kubectl --namespace topograph port-forward $POD_NAME 8080:$CONTAINER_PORT
 
         NOTE: node-data-broker applies node annotations once when each broker pod starts.
-
-        Topology label families use closest-first numeric levels:
-          fabric:      network.topology.nvidia.com/level-N
-          accelerated: accelerated.topology.nvidia.com/level-N
   2: |
     apiVersion: v1
     data:
@@ -2333,8 +2305,6 @@ renders values.k8s.ib-example.yaml:
           containers:
             - args:
                 - -v=3
-                - -k8s-topology-key-fabric-prefix=network.topology.nvidia.com/level-
-                - -k8s-topology-key-accelerated-prefix=accelerated.topology.nvidia.com/level-
               command:
                 - /usr/local/bin/topograph
               env:
@@ -2855,10 +2825,6 @@ renders values.slinky.block-example.yaml:
         kubectl --namespace topograph port-forward $POD_NAME 8080:$CONTAINER_PORT
 
         NOTE: node-data-broker applies node annotations once when each broker pod starts.
-
-        Topology label families use closest-first numeric levels:
-          fabric:      network.topology.nvidia.com/level-N
-          accelerated: accelerated.topology.nvidia.com/level-N
   2: |
     apiVersion: v1
     data:
@@ -2910,8 +2876,6 @@ renders values.slinky.block-example.yaml:
           containers:
             - args:
                 - -v=3
-                - -k8s-topology-key-fabric-prefix=network.topology.nvidia.com/level-
-                - -k8s-topology-key-accelerated-prefix=accelerated.topology.nvidia.com/level-
               command:
                 - /usr/local/bin/topograph
               env:
@@ -3436,10 +3400,6 @@ renders values.slinky.partition-example.yaml:
         kubectl --namespace topograph port-forward $POD_NAME 8080:$CONTAINER_PORT
 
         NOTE: node-data-broker applies node annotations once when each broker pod starts.
-
-        Topology label families use closest-first numeric levels:
-          fabric:      network.topology.nvidia.com/level-N
-          accelerated: accelerated.topology.nvidia.com/level-N
   2: |
     apiVersion: v1
     data:
@@ -3491,8 +3451,6 @@ renders values.slinky.partition-example.yaml:
           containers:
             - args:
                 - -v=3
-                - -k8s-topology-key-fabric-prefix=network.topology.nvidia.com/level-
-                - -k8s-topology-key-accelerated-prefix=accelerated.topology.nvidia.com/level-
               command:
                 - /usr/local/bin/topograph
               env:
@@ -4034,10 +3992,6 @@ renders values.slinky.tree-example.yaml:
         kubectl --namespace topograph port-forward $POD_NAME 8080:$CONTAINER_PORT
 
         NOTE: node-data-broker applies node annotations once when each broker pod starts.
-
-        Topology label families use closest-first numeric levels:
-          fabric:      network.topology.nvidia.com/level-N
-          accelerated: accelerated.topology.nvidia.com/level-N
   2: |
     apiVersion: v1
     data:
@@ -4089,8 +4043,6 @@ renders values.slinky.tree-example.yaml:
           containers:
             - args:
                 - -v=3
-                - -k8s-topology-key-fabric-prefix=network.topology.nvidia.com/level-
-                - -k8s-topology-key-accelerated-prefix=accelerated.topology.nvidia.com/level-
               command:
                 - /usr/local/bin/topograph
               env:

charts/topograph/tests/deployment_test.yaml

@@ -272,25 +272,14 @@ tests:
             name: NODE_DATA_BROKER_NAMESPACE
         template: templates/deployment.yaml
 
-  - it: passes the verbosity flag and topology key args
+  - it: passes the verbosity flag
     set:
       verbosity: 5
-      topologyNodeLabels:
-        fabricPrefix: example.com/fabric-
-        acceleratedPrefix: example.com/accelerated-
     asserts:
       - contains:
           path: spec.template.spec.containers[0].args
           content: -v=5
         template: templates/deployment.yaml
-      - contains:
-          path: spec.template.spec.containers[0].args
-          content: -k8s-topology-key-fabric-prefix=example.com/fabric-
-        template: templates/deployment.yaml
-      - contains:
-          path: spec.template.spec.containers[0].args
-          content: -k8s-topology-key-accelerated-prefix=example.com/accelerated-
-        template: templates/deployment.yaml
 
   - it: renders user-supplied environment variables, initContainers, and lifecycle
     set:

charts/topograph/values.schema.json

@@ -50,7 +50,7 @@
         },
         "params": {
           "type": "object",
-          "description": "Engine-specific parameters. For slinky, useGpuCliqueLabel=true reads nvidia.com/gpu.clique as the topology/block domain source."
+          "description": "Engine-specific parameters. The k8s engine accepts closest-first fabricLabels and narrowest-first acceleratedLabels arrays. Each omitted array defaults every discovered entry in its family; when provided, additional entries in that family are omitted. For slinky, useGpuCliqueLabel=true reads nvidia.com/gpu.clique as the topology/block domain source."
         }
       },
       "required": ["name"]
@@ -119,7 +119,6 @@
     "initContainers": { "type": "array" },
     "lifecycle": { "type": "object" },
     "config": { "type": "object" },
-    "topologyNodeLabels": { "type": "object" },
     "podAnnotations": { "type": "object" },
     "podLabels": { "type": "object" },
     "podSecurityContext": { "type": "object" },

charts/topograph/values.yaml

@@ -67,11 +67,6 @@ config:
   # Optional secret with API credentials
   # credentialsSecret:
 
-topologyNodeLabels:
-  # Level numbers are appended to these prefixes. Level 0 is closest to a node.
-  fabricPrefix: network.topology.nvidia.com/level-
-  acceleratedPrefix: accelerated.topology.nvidia.com/level-
-
 podAnnotations: {}
 podLabels: {}
 

cmd/topograph/main.go

@@ -28,17 +28,13 @@ import (
 
 	"github.com/NVIDIA/topograph/internal/version"
 	"github.com/NVIDIA/topograph/pkg/config"
-	"github.com/NVIDIA/topograph/pkg/engines/k8s"
 	"github.com/NVIDIA/topograph/pkg/server"
 )
 
 func main() {
 	var cfg string
-	var fabricLabelPrefix, acceleratedLabelPrefix string
 	var ver bool
 	flag.StringVar(&cfg, "c", "/etc/topograph/topograph-config.yaml", "config file")
-	flag.StringVar(&fabricLabelPrefix, "k8s-topology-key-fabric-prefix", k8s.DefaultFabricLabelPrefix, "K8s node-label prefix for fabric levels")
-	flag.StringVar(&acceleratedLabelPrefix, "k8s-topology-key-accelerated-prefix", k8s.DefaultAcceleratedLabelPrefix, "K8s node-label prefix for accelerated-network levels")
 	flag.BoolVar(&ver, "version", false, "show the version")
 
 	klog.InitFlags(nil)
@@ -50,8 +46,6 @@ func main() {
 		os.Exit(0)
 	}
 
-	k8s.InitLabels(fabricLabelPrefix, acceleratedLabelPrefix)
-
 	if err := mainInternal(cfg); err != nil {
 		klog.Error(err.Error())
 		os.Exit(1)