Component
core/ai-model/vllm
vLLM running in-cluster as a self-hosted inference server. A Recipe synthesizes the runtime directly: a weights-pull initContainer resolves the model pointer into a model-cache volume the serving container mounts, and a ClusterIP Service fronts the OpenAI-compatible API. The model value is a deterministic pointer resolved at pod start, never fetched at compile time. Exports its host, port, endpoint, model, and protocol to consumers via the uses topology, so an agent reaches the model without hardcoding an endpoint. Each realization also publishes where Prometheus finds the server's metrics: one address as `metrics_endpoint`, or, where one address cannot reach every replica, the Service or the pod labels a scraper discovers them by. On KServe the cluster's serving runtime decides the metrics port and path, so nothing is published there. On a cluster running the Prometheus Operator, `core/prometheus-exporter` attached to the synthesized Kubernetes realization, its port set to `http`, emits a ServiceMonitor for it instead.
Herkunft
- Inhalts-Digest
- sha256:b13caeae34054f971912cfb864d53c9efbe299e59dbd5280ed7ce312e50fed52
- Veröffentlicht
- 2026-11-10T10:09:41Z
- Signaturidentität
- https://token.actions.githubusercontent.com/https://github.com/aggrega-ai/aggrega-hub/.github/workflows/catalog-publish-sign.yml@refs/heads/main
Signatur
- Offiziellapplication/vnd.dev.sigstore.bundle.v0.3+json
https://token.actions.githubusercontent.com/https://github.com/aggrega-ai/aggrega-hub/.github/workflows/catalog-publish-sign.yml@refs/heads/main
sha256:50a6d28f11d3c84b4ec7b3b89bbdecc5d9f1523323fda172c2340fbd633553ce
Referenz zu dieser Version
oci://hub.aggrega.ai/component/core/ai-model/vllm:1.0.0Entwicklerparameter
Werte, die der Entwickler im Aggregate festlegt.
Erweitert BaseComponents#/schemas/BaseModelParameter
| Name | Typ | Standardwert | Erforderlich | Beschreibung |
|---|---|---|---|---|
| served_model_name | string | — | nein | — |
Dimensionierung durch den Operator
Infrastrukturdimensionierung, die der Operator pro Umgebung anpasst.
| Name | Typ | Standardwert | Erforderlich | Beschreibung |
|---|---|---|---|---|
| port | integer (min 1024, max 65535) | 8000 | nein | Server listen port for the model API. Most operators should omit this and let the Recipe assign the default (8000). |
| engine_version | string | v0.6.3 | nein | vLLM runtime image tag. Selects the image tag on the synthesized serving and weights-pull containers. |
| replicas | integer (min 1) | 1 | nein | Number of serving replicas. Each replica holds its own model-cache volume and pulls the weights independently. |
| min_replicas | integer (min 1) | — | nein | Lower bound for the replica count when horizontal autoscaling is configured. Declared ahead of rendering — the synth recipes serve a fixed replica count and do not yet emit an autoscaler; set it to record the intended lower bound. |
| max_replicas | integer (min 1) | — | nein | Upper bound for the replica count when horizontal autoscaling is configured. Declared ahead of rendering — the synth recipes serve a fixed replica count and do not yet emit an autoscaler; set it to record the intended upper bound. |
| compute | object | — | nein | — |
| model_cache | object | — | nein | Per-replica model-weights cache volume. Sized for the pulled model; each serving replica gets its own claim. |
| model_artifact | string | — | nein | Optional OCI reference to the model weights, an alternative weights source to the Hugging Face model id in values.model. A deterministic pointer the runtime resolves at pod start; the compiler never fetches it. Declared ahead of rendering — the synth recipes resolve weights from values.model and do not yet consume this pointer. |
| control_plane_present | boolean | false | nein | Operator affirmation that the model-serving control plane (the KServe or KubeAI controller and its custom-resource definitions) is installed in the target cluster. A controller-driven realization halts the build unless this is set true, so a serving custom resource is never emitted as an orphan into a cluster that cannot reconcile it. The direct synthesis recipes ignore this flag. |
| overrides | object (open) | — | nein | Provider-specific overrides. Accepts any nested structure. Each Recipe documents which keys it reads. |
| gpu | object | — | nein | Optional GPU scheduling for the serving container. When set, the Recipe requests GPUs as the extended resource nvidia.com/gpu and, when a runtime class is named, schedules the pod under it. Requires a GPU-capable cluster and, for runtime_class, a matching RuntimeClass configured in the cluster. Omit the whole object to serve on CPU. |
Ausgabevertrag
Werte, die diese Komponente über die uses-Topologie an Konsumenten veröffentlicht.
| Name | Typ | Standardwert | Erforderlich | Beschreibung |
|---|---|---|---|---|
| host | string | — | ja | In-cluster DNS name of the Service fronting the inference server. Which Service that is follows the realization the Manifest selected, so a consumer reads this rather than deriving a name. |
| port | integer | — | ja | TCP port that Service accepts model requests on. A controller-driven realization fronts the server on its own port, so a consumer reads this rather than assuming a default. |
| endpoint | string | — | ja | Full base URL of the OpenAI-compatible API the server serves, with the host, the port and the API path already composed. A consumer hands it to a client library as its base URL. |
| model | string | — | ja | Name a request must ask this server for. It is the served name the realization registered, which is not always the weights id the Aggregate declared. |
| protocol | string | — | nein | URL scheme to reach `host` and `port` over, for a consumer composing its own address rather than reading `endpoint`. |
| metrics_endpoint | string | — | nein | The full address Prometheus scrapes this server's metrics at — scheme, host, port and path, `http://<host>:<port>/metrics`, on the port the API is served on. Empty where one address cannot reach every replica, which then publishes the service or pod facts instead, and on KServe, which publishes none. |
| metrics_service | string | — | nein | Name of the Service every replica's metrics sit behind, for a scraper that discovers each endpoint of that Service rather than scraping one address. Set by the Kubernetes synthesis while infra.replicas is above one, beside `metrics_service_port` and `metrics_path`; empty otherwise. |
| metrics_service_port | string | — | nein | Name of the port of `metrics_service` that carries the metrics, `http`. Set and empty together with `metrics_service`. |
| metrics_pod_selector | string | — | nein | Kubernetes label selector matching every pod that serves this server's metrics, for a scraper that discovers those pods directly because no Service of the realization carries them. Set by the KubeAI realization, beside `metrics_pod_port` and `metrics_path`; empty otherwise. |
| metrics_pod_port | string | — | nein | Name of the container port on those pods that carries the metrics, `http`. Set and empty together with `metrics_pod_selector`. |
| metrics_path | string | — | nein | Path the metrics are served at, `/metrics`, published beside the service or the pod facts. Empty where the server publishes `metrics_endpoint`, which carries its own path. |
uses-Sockets
Die Abhängigkeits-Sockets, über die sich diese Komponente verbindet. Das Ziel jeder über einen Socket verdrahteten Verbindung muss dessen Label-Selektor erfüllen, und jede Verbindung kann in den Slots, die der Socket deklariert, eine Anbindung tragen. Einen erforderlichen Slot muss jede Verbindung füllen, die das Aggregate oder das Release über den Socket verdrahtet.
Keine uses-Sockets deklariert.
Anbindungs-Slots
Benannte Bindungspunkte, an denen ein Feature an diese Komponente angebunden wird. Der Selektor eines Slots schränkt ein, welches Feature ihn füllen darf, und seine inject-Liste benennt, wie das Feature sein Material liefern darf. Ein erforderlicher Slot muss auf jeder Instanz gebunden sein, durch eine Anbindung an der Instanz oder in einem darüberliegenden Geltungsbereich, und ein sensibler Slot ist für den Umgang mit geheimem Material deklariert. Ein Feature kann auch unter einem Namen angebunden werden, den kein Slot deklariert.
| Slot | Selektor | Erforderlich | Injektion | Sensibel | Beschreibung |
|---|---|---|---|---|---|
| model_pull_token | feature.aggrega.ai/kind=secret | nein | env | ja | Optional model-registry pull credential for gated model weights. When attached, the weights-pull initContainer receives it as an environment variable; unattached, weights are pulled anonymously. |
| api_key | feature.aggrega.ai/kind=secret | nein | env | ja | Optional key the server requires of every caller. When attached, the Recipe delivers it as VLLM_API_KEY, the variable vLLM reads at startup, and each request under /v1 must then carry it as a bearer token in the Authorization header. The server checks this one key, so every caller must present the same key. Unattached, the server requires none. |
Zulassungsliste für Selbstinjektion
env
Beziehungen
Typisierte Kanten zu und von anderen veröffentlichten Entitäten. Eine Referenz ohne veröffentlichtes Ziel wird als unterbrochene Kante statt als Link dargestellt.
Verwendet
- $referweitert Schema (values)
BaseComponents#/schemas/BaseModelParameter
Verwendet von
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «model»
verwendet Socket «target_model»
verwendet Socket «extraction_model»
verwendet Socket «embedding_model»
verwendet Socket «models»
verwendet Socket «models»
verwendet Socket «models»
verwendet Socket «models»
verwendet Socket «models»
verwendet Socket «models»
verwendet Socket «models»
verwendet Socket «upstream»
verwendet Socket «models»
verwendet Socket «forward_target»
verwendet Socket «scrape_target»
verwendet Socket «store»
verwendet Socket «content_safety»
gilt für
gilt für
gilt für
gilt für
gilt für
gilt für
gilt für
implementiert
implementiert
implementiert
implementiert