Component
core/gateway/azure-api-management
The AI gateway of Azure API Management, a hosted service: a caller sends its model requests to an API on an API Management instance, which forwards them to the model API behind it. Nothing runs in the Release and the compiler never calls Azure. It publishes the API's address and the model a caller asks for, which the gateway forwards unchecked. On terraform its Recipe adds the API, its backend and its policy to an instance that already exists, and is the one that can carry controls: reach for it where the instance exists and the Release is to configure the API. On helm and compose its Recipes reference an API that already exists and carry no Feature: reach for those where the API is configured outside the Release. On terraform it realizes core/semantic-caching, core/period-budget, core/prompt-injection-shield and core/retry-policy in the API's policy, and halts core/audit-logging by name. Not realized: core/response-validation, core/target-selection, core/data-masking, core/concurrency-limit and core/fallback-chain. A caller presents its own subscription key; the one credential the gateway holds is the optional provider_key, written as a reference to a Key Vault secret. What the gateway does with a request is Microsoft's documented behaviour and no build observes it. It carries no model.aggrega.ai/api label, so an agent's model socket does not take it.
Procedencia
- Digest del contenido
- sha256:9b57a813e0a17aeb85dd2bf16c049099a481eb135c41c8283efd3bd1ba78ab9a
- Publicado
- 2026-11-10T10:11:41Z
- Identidad de firma
- https://token.actions.githubusercontent.com/https://github.com/aggrega-ai/aggrega-hub/.github/workflows/catalog-publish-sign.yml@refs/heads/main
Firma
- Oficialapplication/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:aba359c7ff0006843e2266a3230daa2bc4777513ae0277a80a8849f6d252e1bb
Referencia en esta versión
oci://hub.aggrega.ai/component/core/gateway/azure-api-management:1.0.0Parámetros del desarrollador
Valores que el desarrollador establece en el Aggregate.
Extiende BaseComponents#/schemas/BaseComponentParameter
| Nombre | Tipo | Predeterminado | Obligatorio | Descripción |
|---|---|---|---|---|
| model | string (min 1) | — | sí | The model a caller asks the gateway for, written as the model API behind it expects it: a deployment name on Azure OpenAI, or the provider's own model name, such as gpt-4o-mini. Published as model; the gateway forwards it without checking it. |
Dimensionamiento del operador
Dimensionamiento de la infraestructura que el operador ajusta para cada entorno.
| Nombre | Tipo | Predeterminado | Obligatorio | Descripción |
|---|---|---|---|---|
| api_management_name | string (min 1, max 50) | — | sí | The name of the API Management instance the API is served on. The terraform Recipe reads this instance and never creates it; the helm and compose Recipes compose the address from it. |
| api_path | string (min 1, max 400) | — | sí | The path the API is served under at the gateway, unique within the instance, such as models or teams/chat. The terraform Recipe creates the API at it and names the API, its backend and its named value after it, with each slash as a hyphen; the helm and compose Recipes reference the API already served there. |
| gateway_url | string | — | no | The scheme and host the instance's gateway answers on, with no path, such as https://api.contoso.com for an instance reached under its own domain. Unset, the address is https://<api_management_name>.azure-api.net, the host Azure gives the instance. |
| resource_group_name | string (min 1) | — | no | The resource group the API Management instance sits in. Read by the terraform Recipe only, which halts the build when it is unset. |
| backend_url | string | — | no | The base address of the model API the gateway forwards to, with no trailing slash: the gateway appends /chat/completions. For Azure OpenAI, https://<resource>.openai.azure.com/openai/v1. Read by the terraform Recipe only, which halts the build when it is unset. |
| backend_authentication | string (managed_identity | provider_key | none) | — | no | How the gateway authenticates to the model API. managed_identity: a token for https://cognitiveservices.azure.com obtained with the instance's system-assigned identity and sent as a bearer token, which Azure OpenAI and other Microsoft Foundry models accept once that identity holds a role on them; the gateway does not check where it sends the token, so choose it only for such a backend. provider_key: the key held in the Key Vault secret provider_key names, sent in the header infra.provider_key_header names. none: nothing is added, for a model API that takes no credential. Read by the terraform Recipe only, which halts the build when it is unset. |
| provider_key_header | string | api-key | no | The header the key provider_key names is sent in. api-key, the default, is the header Azure OpenAI reads; a model API taking a bearer token reads Authorization. Read by the terraform Recipe only, while infra.backend_authentication is provider_key. |
| subscription_required | boolean | true | no | Whether the API refuses a request that presents no API Management subscription key. Off, anyone holding the address can send model requests through the gateway. Written by the terraform Recipe only. |
| content_safety_backend_id | string | — | no | The id of the API Management backend for the instance's Azure AI Content Safety resource, at https://<name>.cognitiveservices.azure.com, which the instance reaches with its managed identity holding the Cognitive Services User role there. Read by the terraform Recipe only, which halts the build when core/prompt-injection-shield is attached and it is unset; the backend is configured on the instance outside the Release. |
| cache_serves_vector_search | boolean | false | no | Affirms that the external cache the API Management instance uses is a Redis cache serving the vector search of the Redis query engine, RediSearch, which Azure's semantic cache looks prompts up with. The cache is configured on the instance outside the Release and the compiler never reads it, so it asks rather than checks. Read by the terraform Recipe only, which halts the build when core/semantic-caching is attached and this is left unaffirmed; with no such attachment it changes nothing rendered. |
Contrato de salida
Valores que este componente publica para los consumidores a través de la topología uses.
| Nombre | Tipo | Predeterminado | Obligatorio | Descripción |
|---|---|---|---|---|
| endpoint | string | — | sí | The address of the API, <gateway_url>/<api_path>, with gateway_url defaulting to https://<api_management_name>.azure-api.net. A client of the OpenAI chat completions API uses it as its base address and presents its subscription key in the Ocp-Apim-Subscription-Key header. |
| model | string | — | sí | The model a caller asks for, carried through from values.model. |
Sockets uses
Los sockets de dependencia a través de los cuales se conecta este componente. El destino de cada conexión cableada a través de un socket debe satisfacer su selector de etiquetas, y cada conexión puede llevar un acoplamiento en los slots que declara el socket. Cada conexión que el Aggregate o la Release cablea a través del socket debe llenar los slots obligatorios.
No se declararon sockets uses.
Slots de acoplamiento
Puntos de enlace con nombre en los que una Feature se acopla a este componente. El selector de un slot restringe qué Feature puede llenarlo y su lista inject indica cómo puede entregarse el material de la Feature. Un slot obligatorio debe estar enlazado en cada instancia, mediante un acoplamiento en la instancia o en un ámbito superior, y un slot sensible se declara para manejar material secreto. Una Feature también puede acoplarse con un nombre que ningún slot declara.
| Slot | Selector | Obligatorio | Inyección | Sensible | Descripción |
|---|---|---|---|---|---|
| provider_key | feature.aggrega.ai/kind=secret | no | env | sí | Optional key the gateway presents to the model API behind it, read only by the terraform Recipe and only while infra.backend_authentication is provider_key. The key itself stays in Azure Key Vault: the attachment delivers the identifier of the Key Vault secret holding it, https://<vault>.vault.azure.net/secrets/<name>, as core/secret-value bound to terraform/variable-secret does, and the Recipe writes that identifier into a secret named value the instance reads with its managed identity, which needs read access to that secret. The build halts on a delivered value that is not such an identifier, so a key written in place of one never reaches a rendered file. The secret holds the header's whole value, as the header named by infra.provider_key_header carries it: for a bearer token, Bearer followed by the key. |
Lista de permitidos de autoinyección
No se declararon canales de autoinyección.
Relaciones
Aristas tipadas hacia y desde otras entidades publicadas. Una referencia sin un destino publicado se muestra como una arista rota en lugar de como un enlace.
Usa
- $refextiende el esquema (values)
BaseComponents#/schemas/BaseComponentParameter
Usado por
se aplica a
se aplica a
se aplica a
se aplica a
se aplica a
se aplica a
se aplica a
se aplica a
implementa
implementa
implementa