From d1ef99abc52d80bf94b8c389bae5b543aaa8472d Mon Sep 17 00:00:00 2001 From: Marcus Souza Date: Mon, 21 Sep 2026 20:18:51 -0300 Subject: [PATCH 1/2] fix(cache): align the overview's scope section with the other product overviews The Cache overview was the only one of the six product overviews built on this template that named its third slot something else: "What Cache covers" where AI Inference, WAF, Functions, Application Accelerator, and Image Processor all say "Scope and limits". Same slot, same kind, a different name to retrieve it by. Rename the heading in both languages, and add the Billing bullet the section was missing. Cache is billed on Purges, the one metric its pricing section carries; data transfer is listed there under Application Accelerator, so the bullet says so rather than letting "billed on Purges" read as "delivery is free". The object and path headings are left alone: those two slots were never uniform across the six pages, and Cache's already match Application Accelerator's. --- .../reference/build/edge-application/edge-caching.mdx | 3 ++- .../referencia/build/edge-application/edge-caching.mdx | 3 ++- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/content/docs/en/pages/main-menu/reference/build/edge-application/edge-caching.mdx b/src/content/docs/en/pages/main-menu/reference/build/edge-application/edge-caching.mdx index 375504ad62..f19c2f7f7b 100644 --- a/src/content/docs/en/pages/main-menu/reference/build/edge-application/edge-caching.mdx +++ b/src/content/docs/en/pages/main-menu/reference/build/edge-application/edge-caching.mdx @@ -76,7 +76,7 @@ For the mechanisms behind each step, refer to [How Cache works](/en/documentatio --- -## What Cache covers +## Scope and limits - **Interfaces.** A cache setting is created and managed in [Azion Console](https://console.azion.com/), through the Azion API v4, the [Azion CLI](/en/documentation/devtools/cli/), `azion.config.js`, [Azion Lib](/en/documentation/devtools/azion-lib/application/), and [Terraform](/en/documentation/devtools/terraform/applications/). - **Methods.** `GET` and `HEAD` responses are cached natively. `POST` and `OPTIONS` responses are cached with [Application Accelerator](/en/documentation/build/application-accelerator/), which also lifts the TTL floor below 60 seconds and turns on cache variation by query string, cookie, and device group. @@ -86,6 +86,7 @@ For the mechanisms behind each step, refer to [How Cache works](/en/documentatio - **Bounds.** The cache TTL runs from 60 seconds, or 0 with Application Accelerator, up to 31,536,000 seconds. A single object is held up to 10 GB, and large files are fragmented at 1,024 kB. Every ceiling, and the purges each plan includes, are on [Cache limits](/en/documentation/build/cache/limits/). - **What it does not do.** Cache stores responses your origin or your application produces; it does not run code. A function that reads and writes cached entries from inside its own code uses the [Cache API](/en/documentation/devtools/runtime/api-reference/cache/) of [Functions](/en/documentation/build/functions/) instead. - **Watching it.** [Real-Time Metrics](/en/documentation/platform/real-time-metrics/) carries a Cache dataset and a Tiered Cache dataset, and a single response reports its own status through a debug header. +- **Billing.** Cache is billed on Purges, and the pricing page lists data transfer under [Application Accelerator](/en/documentation/build/application-accelerator/). For the rate past the amount a plan includes, refer to [Pricing](/en/documentation/fundamentals/pricing/#cache). --- diff --git a/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/edge-caching.mdx b/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/edge-caching.mdx index a36955666e..4c39d13255 100644 --- a/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/edge-caching.mdx +++ b/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/edge-caching.mdx @@ -76,7 +76,7 @@ Para os mecanismos por trás de cada passo, consulte [Como o Cache funciona](/pt --- -## O que o Cache cobre +## Escopo e limites - **Interfaces.** Um cache setting é criado e gerenciado no [Azion Console](https://console.azion.com/), pela Azion API v4, pela [Azion CLI](/pt-br/documentacao/devtools/cli/), pelo `azion.config.js`, pela [Azion Lib](/en/documentation/devtools/azion-lib/application/) e pelo [Terraform](/pt-br/documentacao/devtools/terraform/applications/). - **Métodos.** Respostas `GET` e `HEAD` são colocadas em cache nativamente. Respostas `POST` e `OPTIONS` são colocadas em cache com o [Application Accelerator](/pt-br/documentacao/build/application-accelerator/), que também eleva o piso do TTL abaixo de 60 segundos e ativa a variação de cache por query string, cookie e device group. @@ -86,6 +86,7 @@ Para os mecanismos por trás de cada passo, consulte [Como o Cache funciona](/pt - **Limites.** O TTL de cache vai de 60 segundos, ou 0 com o Application Accelerator, até 31.536.000 segundos. Um objeto é mantido até 10 GB, e arquivos grandes são fragmentados a cada 1.024 kB. Cada limite, e os purges que cada plano inclui, estão em [Limites do Cache](/pt-br/documentacao/build/cache/limites/). - **O que ele não faz.** O Cache armazena respostas que a sua origem ou a sua aplicação produzem; ele não executa código. Uma função que lê e escreve entradas de cache a partir do seu próprio código usa a [Cache API](/pt-br/documentacao/devtools/runtime/api-reference/cache/) do [Functions](/pt-br/documentacao/build/functions/). - **Acompanhar.** O [Real-Time Metrics](/pt-br/documentacao/plataforma/real-time-metrics/) carrega um dataset de Cache e um de Tiered Cache, e uma resposta informa o seu próprio status por um header de debug. +- **Cobrança.** O Cache é cobrado por Purges, e a página de preços lista a transferência de dados sob o [Application Accelerator](/pt-br/documentacao/build/application-accelerator/). Para a taxa acima da quantidade que um plano inclui, consulte [Preços](/pt-br/documentacao/fundamentos/precos/#cache). --- From 743a7aac453ef417127439e52f4b8da8173a7701 Mon Sep 17 00:00:00 2001 From: Marcus Souza Date: Mon, 21 Sep 2026 20:25:12 -0300 Subject: [PATCH 2/2] fix(docs): normalize the scope-section bullet labels on the Cache and Image Processor overviews Four of the six product overviews built on this template write their scope bullets as "**Interfaces**: the module...". Cache and Image Processor wrote them as "**Interfaces.** The module...", so the same slot read two ways across the set. Convert both pages in both languages to the colon form, lowercasing the lead word where it is an ordinary one and leaving product names, acronyms, and code as they are, which is the convention the other four already follow. Cache's observability bullet was also named "Watching it" where AI Inference, WAF, and Functions all say "Observability". Rename it, and its Portuguese "Acompanhar" to "Observabilidade". The Portuguese Cache overview linked Azion Lib at an /en/ URL. Point it at the Portuguese page. That page's permalink reads /documentacao/produtos/azion-lib/ application/ rather than the /devtools/ its thirteen siblings and all fourteen English pages use, but the produtos URL is a live redirect destination in cicd/massive-redirect/pt-br.json, so the link follows the URL that resolves today and the permalink is left for a URL-ledger decision. --- .../build/edge-application/edge-caching.mdx | 18 +++++++++--------- .../build/edge-application/image-processor.mdx | 12 ++++++------ .../build/edge-application/edge-caching.mdx | 18 +++++++++--------- .../build/edge-application/image-processor.mdx | 12 ++++++------ 4 files changed, 30 insertions(+), 30 deletions(-) diff --git a/src/content/docs/en/pages/main-menu/reference/build/edge-application/edge-caching.mdx b/src/content/docs/en/pages/main-menu/reference/build/edge-application/edge-caching.mdx index f19c2f7f7b..04b17da8d1 100644 --- a/src/content/docs/en/pages/main-menu/reference/build/edge-application/edge-caching.mdx +++ b/src/content/docs/en/pages/main-menu/reference/build/edge-application/edge-caching.mdx @@ -78,15 +78,15 @@ For the mechanisms behind each step, refer to [How Cache works](/en/documentatio ## Scope and limits -- **Interfaces.** A cache setting is created and managed in [Azion Console](https://console.azion.com/), through the Azion API v4, the [Azion CLI](/en/documentation/devtools/cli/), `azion.config.js`, [Azion Lib](/en/documentation/devtools/azion-lib/application/), and [Terraform](/en/documentation/devtools/terraform/applications/). -- **Methods.** `GET` and `HEAD` responses are cached natively. `POST` and `OPTIONS` responses are cached with [Application Accelerator](/en/documentation/build/application-accelerator/), which also lifts the TTL floor below 60 seconds and turns on cache variation by query string, cookie, and device group. -- **One URL, several copies.** A setting can store one copy per query-string argument, cookie value, device group, request method, image format, or byte range of a large file. Each is a separate key, so each is purged separately. The formats are on [Cache keys](/en/documentation/build/cache/cache-keys/). -- **A second layer.** [Tiered Cache](/en/documentation/build/cache/tiered-cache/) answers a miss in the first layer before the origin is asked, and is turned on per cache setting. -- **Removing a copy.** [Real-Time Purge](/en/documentation/build/cache/real-time-purge/) removes an object by URL, by cache key, or by a wildcard expression, from either layer, before its TTL ends. -- **Bounds.** The cache TTL runs from 60 seconds, or 0 with Application Accelerator, up to 31,536,000 seconds. A single object is held up to 10 GB, and large files are fragmented at 1,024 kB. Every ceiling, and the purges each plan includes, are on [Cache limits](/en/documentation/build/cache/limits/). -- **What it does not do.** Cache stores responses your origin or your application produces; it does not run code. A function that reads and writes cached entries from inside its own code uses the [Cache API](/en/documentation/devtools/runtime/api-reference/cache/) of [Functions](/en/documentation/build/functions/) instead. -- **Watching it.** [Real-Time Metrics](/en/documentation/platform/real-time-metrics/) carries a Cache dataset and a Tiered Cache dataset, and a single response reports its own status through a debug header. -- **Billing.** Cache is billed on Purges, and the pricing page lists data transfer under [Application Accelerator](/en/documentation/build/application-accelerator/). For the rate past the amount a plan includes, refer to [Pricing](/en/documentation/fundamentals/pricing/#cache). +- **Interfaces**: a cache setting is created and managed in [Azion Console](https://console.azion.com/), through the Azion API v4, the [Azion CLI](/en/documentation/devtools/cli/), `azion.config.js`, [Azion Lib](/en/documentation/devtools/azion-lib/application/), and [Terraform](/en/documentation/devtools/terraform/applications/). +- **Methods**: `GET` and `HEAD` responses are cached natively. `POST` and `OPTIONS` responses are cached with [Application Accelerator](/en/documentation/build/application-accelerator/), which also lifts the TTL floor below 60 seconds and turns on cache variation by query string, cookie, and device group. +- **One URL, several copies**: a setting can store one copy per query-string argument, cookie value, device group, request method, image format, or byte range of a large file. Each is a separate key, so each is purged separately. The formats are on [Cache keys](/en/documentation/build/cache/cache-keys/). +- **A second layer**: [Tiered Cache](/en/documentation/build/cache/tiered-cache/) answers a miss in the first layer before the origin is asked, and is turned on per cache setting. +- **Removing a copy**: [Real-Time Purge](/en/documentation/build/cache/real-time-purge/) removes an object by URL, by cache key, or by a wildcard expression, from either layer, before its TTL ends. +- **Bounds**: the cache TTL runs from 60 seconds, or 0 with Application Accelerator, up to 31,536,000 seconds. A single object is held up to 10 GB, and large files are fragmented at 1,024 kB. Every ceiling, and the purges each plan includes, are on [Cache limits](/en/documentation/build/cache/limits/). +- **What it does not do**: Cache stores responses your origin or your application produces; it does not run code. A function that reads and writes cached entries from inside its own code uses the [Cache API](/en/documentation/devtools/runtime/api-reference/cache/) of [Functions](/en/documentation/build/functions/) instead. +- **Observability**: [Real-Time Metrics](/en/documentation/platform/real-time-metrics/) carries a Cache dataset and a Tiered Cache dataset, and a single response reports its own status through a debug header. +- **Billing**: Cache is billed on Purges, and the pricing page lists data transfer under [Application Accelerator](/en/documentation/build/application-accelerator/). For the rate past the amount a plan includes, refer to [Pricing](/en/documentation/fundamentals/pricing/#cache). --- diff --git a/src/content/docs/en/pages/main-menu/reference/build/edge-application/image-processor.mdx b/src/content/docs/en/pages/main-menu/reference/build/edge-application/image-processor.mdx index 0f734929bb..5fbeaf4706 100644 --- a/src/content/docs/en/pages/main-menu/reference/build/edge-application/image-processor.mdx +++ b/src/content/docs/en/pages/main-menu/reference/build/edge-application/image-processor.mdx @@ -65,12 +65,12 @@ For what each step costs and why the cache key matters, refer to [How Image Proc ## Scope and limits -- **Operations.** Resize, crop, rotate, quality, watermark, format conversion, fill, and any combination of them in one string. Each one and the values it accepts are on [URL parameters](/en/documentation/build/image-processor/url-parameters/). -- **Formats.** JPEG, GIF, PNG, BMP, ICO, WEBP, and AVIF. Conversion to WEBP or AVIF needs the request to carry the matching `Accept` header, and BMP is converted automatically according to what the browser supports. -- **Bounds on one image.** A maximum size of 150 MB, a maximum width of 3840 pixels, and a maximum height of 2160 pixels. The monthly volume of processed images is set by the service plan. Both are on [Limits](/en/documentation/build/image-processor/limits/). -- **Interfaces.** The module is turned on from Azion Console, the Azion API, the Azion CLI, and `azion.config.js`. Every field and flag is on [Settings](/en/documentation/build/image-processor/settings/). -- **Caching.** A derived image is cached under a key that includes its `ims` value, and an image served from the cache is not processed again. Varying the cache by a query string requires the [Application Accelerator](/en/documentation/build/application-accelerator/) module. -- **What it does not do.** Image Processor stores nothing and accepts no uploads. The source image stays where it already is, and no derived image becomes an asset on your account. To process images inside a function instead, refer to the [WASM Image Processor](/en/documentation/devtools/azion-lib/wasm-image-processor/) library, which is a different surface. +- **Operations**: resize, crop, rotate, quality, watermark, format conversion, fill, and any combination of them in one string. Each one and the values it accepts are on [URL parameters](/en/documentation/build/image-processor/url-parameters/). +- **Formats**: JPEG, GIF, PNG, BMP, ICO, WEBP, and AVIF. Conversion to WEBP or AVIF needs the request to carry the matching `Accept` header, and BMP is converted automatically according to what the browser supports. +- **Bounds on one image**: a maximum size of 150 MB, a maximum width of 3840 pixels, and a maximum height of 2160 pixels. The monthly volume of processed images is set by the service plan. Both are on [Limits](/en/documentation/build/image-processor/limits/). +- **Interfaces**: the module is turned on from Azion Console, the Azion API, the Azion CLI, and `azion.config.js`. Every field and flag is on [Settings](/en/documentation/build/image-processor/settings/). +- **Caching**: a derived image is cached under a key that includes its `ims` value, and an image served from the cache is not processed again. Varying the cache by a query string requires the [Application Accelerator](/en/documentation/build/application-accelerator/) module. +- **What it does not do**: Image Processor stores nothing and accepts no uploads. The source image stays where it already is, and no derived image becomes an asset on your account. To process images inside a function instead, refer to the [WASM Image Processor](/en/documentation/devtools/azion-lib/wasm-image-processor/) library, which is a different surface. --- diff --git a/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/edge-caching.mdx b/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/edge-caching.mdx index 4c39d13255..4ca4c2e59b 100644 --- a/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/edge-caching.mdx +++ b/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/edge-caching.mdx @@ -78,15 +78,15 @@ Para os mecanismos por trás de cada passo, consulte [Como o Cache funciona](/pt ## Escopo e limites -- **Interfaces.** Um cache setting é criado e gerenciado no [Azion Console](https://console.azion.com/), pela Azion API v4, pela [Azion CLI](/pt-br/documentacao/devtools/cli/), pelo `azion.config.js`, pela [Azion Lib](/en/documentation/devtools/azion-lib/application/) e pelo [Terraform](/pt-br/documentacao/devtools/terraform/applications/). -- **Métodos.** Respostas `GET` e `HEAD` são colocadas em cache nativamente. Respostas `POST` e `OPTIONS` são colocadas em cache com o [Application Accelerator](/pt-br/documentacao/build/application-accelerator/), que também eleva o piso do TTL abaixo de 60 segundos e ativa a variação de cache por query string, cookie e device group. -- **Uma URL, várias cópias.** Um setting pode armazenar uma cópia por argumento de query string, valor de cookie, device group, método da requisição, formato de imagem ou intervalo de bytes de um arquivo grande. Cada uma é uma key separada, então cada uma é purgada separadamente. Os formatos estão em [Cache keys](/pt-br/documentacao/build/cache/cache-keys/). -- **Uma segunda camada.** O [Tiered Cache](/pt-br/documentacao/build/cache/tiered-cache/) responde a um miss na primeira camada antes de a origem ser consultada, e é ativado por cache setting. -- **Remover uma cópia.** O [Real-Time Purge](/pt-br/documentacao/build/cache/real-time-purge/) remove um objeto por URL, por cache key ou por uma expressão com wildcard, de qualquer uma das camadas, antes do fim do seu TTL. -- **Limites.** O TTL de cache vai de 60 segundos, ou 0 com o Application Accelerator, até 31.536.000 segundos. Um objeto é mantido até 10 GB, e arquivos grandes são fragmentados a cada 1.024 kB. Cada limite, e os purges que cada plano inclui, estão em [Limites do Cache](/pt-br/documentacao/build/cache/limites/). -- **O que ele não faz.** O Cache armazena respostas que a sua origem ou a sua aplicação produzem; ele não executa código. Uma função que lê e escreve entradas de cache a partir do seu próprio código usa a [Cache API](/pt-br/documentacao/devtools/runtime/api-reference/cache/) do [Functions](/pt-br/documentacao/build/functions/). -- **Acompanhar.** O [Real-Time Metrics](/pt-br/documentacao/plataforma/real-time-metrics/) carrega um dataset de Cache e um de Tiered Cache, e uma resposta informa o seu próprio status por um header de debug. -- **Cobrança.** O Cache é cobrado por Purges, e a página de preços lista a transferência de dados sob o [Application Accelerator](/pt-br/documentacao/build/application-accelerator/). Para a taxa acima da quantidade que um plano inclui, consulte [Preços](/pt-br/documentacao/fundamentos/precos/#cache). +- **Interfaces**: um cache setting é criado e gerenciado no [Azion Console](https://console.azion.com/), pela Azion API v4, pela [Azion CLI](/pt-br/documentacao/devtools/cli/), pelo `azion.config.js`, pela [Azion Lib](/pt-br/documentacao/produtos/azion-lib/application/) e pelo [Terraform](/pt-br/documentacao/devtools/terraform/applications/). +- **Métodos**: respostas `GET` e `HEAD` são colocadas em cache nativamente. Respostas `POST` e `OPTIONS` são colocadas em cache com o [Application Accelerator](/pt-br/documentacao/build/application-accelerator/), que também eleva o piso do TTL abaixo de 60 segundos e ativa a variação de cache por query string, cookie e device group. +- **Uma URL, várias cópias**: um setting pode armazenar uma cópia por argumento de query string, valor de cookie, device group, método da requisição, formato de imagem ou intervalo de bytes de um arquivo grande. Cada uma é uma key separada, então cada uma é purgada separadamente. Os formatos estão em [Cache keys](/pt-br/documentacao/build/cache/cache-keys/). +- **Uma segunda camada**: o [Tiered Cache](/pt-br/documentacao/build/cache/tiered-cache/) responde a um miss na primeira camada antes de a origem ser consultada, e é ativado por cache setting. +- **Remover uma cópia**: o [Real-Time Purge](/pt-br/documentacao/build/cache/real-time-purge/) remove um objeto por URL, por cache key ou por uma expressão com wildcard, de qualquer uma das camadas, antes do fim do seu TTL. +- **Limites**: o TTL de cache vai de 60 segundos, ou 0 com o Application Accelerator, até 31.536.000 segundos. Um objeto é mantido até 10 GB, e arquivos grandes são fragmentados a cada 1.024 kB. Cada limite, e os purges que cada plano inclui, estão em [Limites do Cache](/pt-br/documentacao/build/cache/limites/). +- **O que ele não faz**: o Cache armazena respostas que a sua origem ou a sua aplicação produzem; ele não executa código. Uma função que lê e escreve entradas de cache a partir do seu próprio código usa a [Cache API](/pt-br/documentacao/devtools/runtime/api-reference/cache/) do [Functions](/pt-br/documentacao/build/functions/). +- **Observabilidade**: o [Real-Time Metrics](/pt-br/documentacao/plataforma/real-time-metrics/) carrega um dataset de Cache e um de Tiered Cache, e uma resposta informa o seu próprio status por um header de debug. +- **Cobrança**: o Cache é cobrado por Purges, e a página de preços lista a transferência de dados sob o [Application Accelerator](/pt-br/documentacao/build/application-accelerator/). Para a taxa acima da quantidade que um plano inclui, consulte [Preços](/pt-br/documentacao/fundamentos/precos/#cache). --- diff --git a/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/image-processor.mdx b/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/image-processor.mdx index eb70064fd7..7669b10312 100644 --- a/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/image-processor.mdx +++ b/src/content/docs/pt-br/pages/menu-principal/referencia/build/edge-application/image-processor.mdx @@ -66,12 +66,12 @@ Para saber o que cada etapa custa e por que a chave de cache importa, consulte [ ## Escopo e limites -- **Operações.** Redimensionamento, corte, rotação, qualidade, marca d'água, conversão de formato, preenchimento, e qualquer combinação deles em uma única string. Cada uma delas e os valores que aceita estão em [Parâmetros de URL](/pt-br/documentacao/build/image-processor/parametros-de-url/). -- **Formatos.** JPEG, GIF, PNG, BMP, ICO, WEBP e AVIF. A conversão para WEBP ou AVIF exige que a requisição carregue o cabeçalho `Accept` correspondente, e o BMP é convertido automaticamente de acordo com o que o navegador aceita. -- **Limites em uma imagem.** Um tamanho máximo de 150 MB, uma largura máxima de 3840 pixels e uma altura máxima de 2160 pixels. O volume mensal de imagens processadas é definido pelo plano de serviço. Ambos estão em [Limites](/pt-br/documentacao/build/image-processor/limites/). -- **Interfaces.** O módulo é ativado a partir do Azion Console, da Azion API, da Azion CLI e do `azion.config.js`. Cada campo e flag está em [Configurações](/pt-br/documentacao/build/image-processor/configuracoes/). -- **Cache.** Uma imagem derivada é armazenada em cache sob uma chave que inclui seu valor `ims`, e uma imagem servida a partir do cache não é processada novamente. Variar o cache por uma query string exige o módulo [Application Accelerator](/pt-br/documentacao/build/application-accelerator/). -- **O que não faz.** O Image Processor não armazena nada e não aceita uploads. A imagem de origem permanece onde já está, e nenhuma imagem derivada se torna um ativo na sua conta. Para processar imagens dentro de uma função, consulte a biblioteca [WASM Image Processor](/pt-br/documentacao/devtools/azion-lib/wasm-image-processor/), que é uma superfície diferente. +- **Operações**: redimensionamento, corte, rotação, qualidade, marca d'água, conversão de formato, preenchimento, e qualquer combinação deles em uma única string. Cada uma delas e os valores que aceita estão em [Parâmetros de URL](/pt-br/documentacao/build/image-processor/parametros-de-url/). +- **Formatos**: JPEG, GIF, PNG, BMP, ICO, WEBP e AVIF. A conversão para WEBP ou AVIF exige que a requisição carregue o cabeçalho `Accept` correspondente, e o BMP é convertido automaticamente de acordo com o que o navegador aceita. +- **Limites em uma imagem**: um tamanho máximo de 150 MB, uma largura máxima de 3840 pixels e uma altura máxima de 2160 pixels. O volume mensal de imagens processadas é definido pelo plano de serviço. Ambos estão em [Limites](/pt-br/documentacao/build/image-processor/limites/). +- **Interfaces**: o módulo é ativado a partir do Azion Console, da Azion API, da Azion CLI e do `azion.config.js`. Cada campo e flag está em [Configurações](/pt-br/documentacao/build/image-processor/configuracoes/). +- **Cache**: uma imagem derivada é armazenada em cache sob uma chave que inclui seu valor `ims`, e uma imagem servida a partir do cache não é processada novamente. Variar o cache por uma query string exige o módulo [Application Accelerator](/pt-br/documentacao/build/application-accelerator/). +- **O que não faz**: o Image Processor não armazena nada e não aceita uploads. A imagem de origem permanece onde já está, e nenhuma imagem derivada se torna um ativo na sua conta. Para processar imagens dentro de uma função, consulte a biblioteca [WASM Image Processor](/pt-br/documentacao/devtools/azion-lib/wasm-image-processor/), que é uma superfície diferente. ---