From ac43ec9de93deb645962f785b1c2877fc3ae7dce Mon Sep 17 00:00:00 2001 From: Kathryn Alexander Date: Fri, 21 Aug 2026 14:11:21 -0400 Subject: [PATCH] microshift AI jtbd --- ...ardware-accelerators-for-ai-workloads.adoc | 10 ++++++++ maps/jobs/customize-serving-environment.adoc | 6 +++++ maps/jobs/expose-and-query-ai-models.adoc | 14 +++++++++++ maps/jobs/install-ai-capabilities.adoc | 10 ++++++++ maps/jobs/monitor-model-performance.adoc | 8 ++++++ maps/jobs/prepare-and-package-ai-models.adoc | 10 ++++++++ .../remove-ai-model-serving-components.adoc | 4 +++ maps/jobs/secure-ai-inference-endpoints.adoc | 6 +++++ maps/jobs/serve-ai-models.adoc | 16 ++++++++++++ maps/jobs/update-running-ai-model.adoc | 4 +++ ...use-alternative-storage-for-ai-models.adoc | 8 ++++++ maps/microshift/ai.adoc | 25 +++++++++++++++++++ maps/microshift/navigation.adoc | 2 ++ ...ardware-accelerators-for-ai-workloads.adoc | 10 ++++++++ ...roshift-customize-serving-environment.adoc | 10 ++++++++ ...microshift-expose-and-query-ai-models.adoc | 10 ++++++++ .../microshift-install-ai-capabilities.adoc | 10 ++++++++ .../microshift-monitor-model-performance.adoc | 10 ++++++++ ...roshift-prepare-and-package-ai-models.adoc | 10 ++++++++ ...ft-remove-ai-model-serving-components.adoc | 15 +++++++++++ modules/microshift-rhoai-create-ns.adoc | 2 +- modules/microshift-rhoai-create-route.adoc | 2 +- .../microshift-rhoai-export-metrics-otel.adoc | 2 +- ...shift-rhoai-get-model-ready-inference.adoc | 2 +- ...oshift-rhoai-get-model-server-metrics.adoc | 2 +- .../microshift-rhoai-inferenceservice-ex.adoc | 2 +- .../microshift-rhoai-model-package-oci.adoc | 2 +- ...roshift-rhoai-model-serving-rt-verify.adoc | 2 +- ...croshift-rhoai-override-kserve-config.adoc | 2 +- modules/microshift-rhoai-query-model.adoc | 2 +- .../microshift-rhoai-servingruntimes-ex.adoc | 2 +- ...croshift-rhoai-verify-model-connected.adoc | 2 +- ...roshift-secure-ai-inference-endpoints.adoc | 10 ++++++++ modules/microshift-serve-ai-models.adoc | 10 ++++++++ .../microshift-update-running-ai-model.adoc | 15 +++++++++++ ...use-alternative-storage-for-ai-models.adoc | 10 ++++++++ 36 files changed, 255 insertions(+), 12 deletions(-) create mode 100644 maps/jobs/configure-hardware-accelerators-for-ai-workloads.adoc create mode 100644 maps/jobs/customize-serving-environment.adoc create mode 100644 maps/jobs/expose-and-query-ai-models.adoc create mode 100644 maps/jobs/install-ai-capabilities.adoc create mode 100644 maps/jobs/monitor-model-performance.adoc create mode 100644 maps/jobs/prepare-and-package-ai-models.adoc create mode 100644 maps/jobs/remove-ai-model-serving-components.adoc create mode 100644 maps/jobs/secure-ai-inference-endpoints.adoc create mode 100644 maps/jobs/serve-ai-models.adoc create mode 100644 maps/jobs/update-running-ai-model.adoc create mode 100644 maps/jobs/use-alternative-storage-for-ai-models.adoc create mode 100644 maps/microshift/ai.adoc create mode 100644 modules/microshift-configure-hardware-accelerators-for-ai-workloads.adoc create mode 100644 modules/microshift-customize-serving-environment.adoc create mode 100644 modules/microshift-expose-and-query-ai-models.adoc create mode 100644 modules/microshift-install-ai-capabilities.adoc create mode 100644 modules/microshift-monitor-model-performance.adoc create mode 100644 modules/microshift-prepare-and-package-ai-models.adoc create mode 100644 modules/microshift-remove-ai-model-serving-components.adoc create mode 100644 modules/microshift-secure-ai-inference-endpoints.adoc create mode 100644 modules/microshift-serve-ai-models.adoc create mode 100644 modules/microshift-update-running-ai-model.adoc create mode 100644 modules/microshift-use-alternative-storage-for-ai-models.adoc diff --git a/maps/jobs/configure-hardware-accelerators-for-ai-workloads.adoc b/maps/jobs/configure-hardware-accelerators-for-ai-workloads.adoc new file mode 100644 index 00000000000..627403a6220 --- /dev/null +++ b/maps/jobs/configure-hardware-accelerators-for-ai-workloads.adoc @@ -0,0 +1,10 @@ +:_mod-docs-content-type: MAP +:context: configure-hardware-accelerators-for-ai-workloads + +include::modules/microshift-configure-hardware-accelerators-for-ai-workloads.adoc[leveloffset=+0,chunk="to-content"] + +include::modules/microshift-rhoai-workflow.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-install.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-inferenceservice-more-options.adoc[leveloffset=+1,toc="no"] diff --git a/maps/jobs/customize-serving-environment.adoc b/maps/jobs/customize-serving-environment.adoc new file mode 100644 index 00000000000..f3e41195957 --- /dev/null +++ b/maps/jobs/customize-serving-environment.adoc @@ -0,0 +1,6 @@ +:_mod-docs-content-type: MAP +:context: customize-serving-environment + +include::modules/microshift-customize-serving-environment.adoc[leveloffset=+0,chunk="to-content"] + +include::modules/microshift-rhoai-override-kserve-config.adoc[leveloffset=+1,toc="no"] diff --git a/maps/jobs/expose-and-query-ai-models.adoc b/maps/jobs/expose-and-query-ai-models.adoc new file mode 100644 index 00000000000..61af91244bf --- /dev/null +++ b/maps/jobs/expose-and-query-ai-models.adoc @@ -0,0 +1,14 @@ +:_mod-docs-content-type: MAP +:context: expose-and-query-ai-models + +include::modules/microshift-expose-and-query-ai-models.adoc[leveloffset=+0,chunk="to-content"] + +include::modules/microshift-rhoai-create-route.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-query-model-con.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-query-model.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-verify-model-connected.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-get-model-ready-inference.adoc[leveloffset=+1,toc="no"] diff --git a/maps/jobs/install-ai-capabilities.adoc b/maps/jobs/install-ai-capabilities.adoc new file mode 100644 index 00000000000..91b3b6b8c27 --- /dev/null +++ b/maps/jobs/install-ai-capabilities.adoc @@ -0,0 +1,10 @@ +:_mod-docs-content-type: MAP +:context: install-ai-capabilities + +include::modules/microshift-install-ai-capabilities.adoc[leveloffset=+0,chunk="to-content"] + +include::modules/microshift-rhoai-con.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-workflow.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-install.adoc[leveloffset=+1,toc="no"] diff --git a/maps/jobs/monitor-model-performance.adoc b/maps/jobs/monitor-model-performance.adoc new file mode 100644 index 00000000000..60f5782308b --- /dev/null +++ b/maps/jobs/monitor-model-performance.adoc @@ -0,0 +1,8 @@ +:_mod-docs-content-type: MAP +:context: monitor-model-performance + +include::modules/microshift-monitor-model-performance.adoc[leveloffset=+0,chunk="to-content"] + +include::modules/microshift-rhoai-get-model-server-metrics.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-export-metrics-otel.adoc[leveloffset=+1,toc="no"] diff --git a/maps/jobs/prepare-and-package-ai-models.adoc b/maps/jobs/prepare-and-package-ai-models.adoc new file mode 100644 index 00000000000..16b843dbb65 --- /dev/null +++ b/maps/jobs/prepare-and-package-ai-models.adoc @@ -0,0 +1,10 @@ +:_mod-docs-content-type: MAP +:context: prepare-and-package-ai-models + +include::modules/microshift-prepare-and-package-ai-models.adoc[leveloffset=+0,chunk="to-content"] + +include::modules/microshift-rhoai-create-ns.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-model-package-oci.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-workflow.adoc[leveloffset=+1,toc="no"] diff --git a/maps/jobs/remove-ai-model-serving-components.adoc b/maps/jobs/remove-ai-model-serving-components.adoc new file mode 100644 index 00000000000..c64bc551ccf --- /dev/null +++ b/maps/jobs/remove-ai-model-serving-components.adoc @@ -0,0 +1,4 @@ +:_mod-docs-content-type: MAP +:context: remove-ai-model-serving-components + +include::modules/microshift-remove-ai-model-serving-components.adoc[leveloffset=+0,chunk="to-content"] diff --git a/maps/jobs/secure-ai-inference-endpoints.adoc b/maps/jobs/secure-ai-inference-endpoints.adoc new file mode 100644 index 00000000000..1589347accd --- /dev/null +++ b/maps/jobs/secure-ai-inference-endpoints.adoc @@ -0,0 +1,6 @@ +:_mod-docs-content-type: MAP +:context: secure-ai-inference-endpoints + +include::modules/microshift-secure-ai-inference-endpoints.adoc[leveloffset=+0,chunk="to-content"] + +include::modules/microshift-rhoai-create-route.adoc[leveloffset=+1,toc="no"] diff --git a/maps/jobs/serve-ai-models.adoc b/maps/jobs/serve-ai-models.adoc new file mode 100644 index 00000000000..f5888c089af --- /dev/null +++ b/maps/jobs/serve-ai-models.adoc @@ -0,0 +1,16 @@ +:_mod-docs-content-type: MAP +:context: serve-ai-models + +include::modules/microshift-serve-ai-models.adoc[leveloffset=+0,chunk="to-content"] + +include::modules/microshift-rhoai-serving-ai-models-con.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-supported-crds.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-supported-mserv-runtimes.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-servingruntimes-ex.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-inferenceservice-ex.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-model-serving-rt-verify.adoc[leveloffset=+1,toc="no"] diff --git a/maps/jobs/update-running-ai-model.adoc b/maps/jobs/update-running-ai-model.adoc new file mode 100644 index 00000000000..51690087a97 --- /dev/null +++ b/maps/jobs/update-running-ai-model.adoc @@ -0,0 +1,4 @@ +:_mod-docs-content-type: MAP +:context: update-running-ai-model + +include::modules/microshift-update-running-ai-model.adoc[leveloffset=+0,chunk="to-content"] diff --git a/maps/jobs/use-alternative-storage-for-ai-models.adoc b/maps/jobs/use-alternative-storage-for-ai-models.adoc new file mode 100644 index 00000000000..30efa810aeb --- /dev/null +++ b/maps/jobs/use-alternative-storage-for-ai-models.adoc @@ -0,0 +1,8 @@ +:_mod-docs-content-type: MAP +:context: use-alternative-storage-for-ai-models + +include::modules/microshift-use-alternative-storage-for-ai-models.adoc[leveloffset=+0,chunk="to-content"] + +include::modules/microshift-rhoai-model-package-oci.adoc[leveloffset=+1,toc="no"] + +include::modules/microshift-rhoai-workflow.adoc[leveloffset=+1,toc="no"] diff --git a/maps/microshift/ai.adoc b/maps/microshift/ai.adoc new file mode 100644 index 00000000000..19d14c0beed --- /dev/null +++ b/maps/microshift/ai.adoc @@ -0,0 +1,25 @@ +:_mod-docs-content-type: MAP + += AI + +include::jobs/install-ai-capabilities.adoc[leveloffset=+1] + +include::jobs/prepare-and-package-ai-models.adoc[leveloffset=+1] + +include::jobs/serve-ai-models.adoc[leveloffset=+1] + +include::jobs/expose-and-query-ai-models.adoc[leveloffset=+1] + +include::jobs/monitor-model-performance.adoc[leveloffset=+1] + +include::jobs/customize-serving-environment.adoc[leveloffset=+1] + +include::jobs/use-alternative-storage-for-ai-models.adoc[leveloffset=+1] + +include::jobs/configure-hardware-accelerators-for-ai-workloads.adoc[leveloffset=+1] + +include::jobs/update-running-ai-model.adoc[leveloffset=+1] + +include::jobs/secure-ai-inference-endpoints.adoc[leveloffset=+1] + +include::jobs/remove-ai-model-serving-components.adoc[leveloffset=+1] diff --git a/maps/microshift/navigation.adoc b/maps/microshift/navigation.adoc index a5fb04de732..b033790f246 100644 --- a/maps/microshift/navigation.adoc +++ b/maps/microshift/navigation.adoc @@ -20,6 +20,8 @@ include::configure.adoc[leveloffset=+1] include::develop.adoc[leveloffset=+1] +include::ai.adoc[leveloffset=+1] + include::upgrade.adoc[leveloffset=+1] include::plan.adoc[leveloffset=+1] diff --git a/modules/microshift-configure-hardware-accelerators-for-ai-workloads.adoc b/modules/microshift-configure-hardware-accelerators-for-ai-workloads.adoc new file mode 100644 index 00000000000..e321c739387 --- /dev/null +++ b/modules/microshift-configure-hardware-accelerators-for-ai-workloads.adoc @@ -0,0 +1,10 @@ +// Module included in the following assemblies: +// +// * maps/jobs/configure-hardware-accelerators-for-ai-workloads.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-configure-hardware-accelerators-for-ai-workloads_{context}"] += Configure hardware accelerators for AI workloads + +[role="_abstract"] +Configure GPU or NPU accelerators and device plugins so your {microshift-short} node can run inference workloads that require specialized hardware. diff --git a/modules/microshift-customize-serving-environment.adoc b/modules/microshift-customize-serving-environment.adoc new file mode 100644 index 00000000000..8478293bfe1 --- /dev/null +++ b/modules/microshift-customize-serving-environment.adoc @@ -0,0 +1,10 @@ +// Module included in the following assemblies: +// +// * maps/jobs/customize-serving-environment.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-customize-serving-environment_{context}"] += Adapt AI model serving to your edge constraints + +[role="_abstract"] +Override the KServe configuration when the default model-serving environment does not meet your edge deployment constraints. diff --git a/modules/microshift-expose-and-query-ai-models.adoc b/modules/microshift-expose-and-query-ai-models.adoc new file mode 100644 index 00000000000..e988dbf72e3 --- /dev/null +++ b/modules/microshift-expose-and-query-ai-models.adoc @@ -0,0 +1,10 @@ +// Module included in the following assemblies: +// +// * maps/jobs/expose-and-query-ai-models.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-expose-and-query-ai-models_{context}"] += Access your AI model from external applications + +[role="_abstract"] +Create a network route and query your running model so you can make external inferences and test model responses. diff --git a/modules/microshift-install-ai-capabilities.adoc b/modules/microshift-install-ai-capabilities.adoc new file mode 100644 index 00000000000..772e07e6ef4 --- /dev/null +++ b/modules/microshift-install-ai-capabilities.adoc @@ -0,0 +1,10 @@ +// Module included in the following assemblies: +// +// * maps/jobs/install-ai-capabilities.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-install-ai-capabilities_{context}"] += Enable AI model serving on {microshift-short} + +[role="_abstract"] +Install the {rhoai-full} RPM on your {microshift-short} node so you can deploy and serve AI models in edge environments. diff --git a/modules/microshift-monitor-model-performance.adoc b/modules/microshift-monitor-model-performance.adoc new file mode 100644 index 00000000000..c6dca17a483 --- /dev/null +++ b/modules/microshift-monitor-model-performance.adoc @@ -0,0 +1,10 @@ +// Module included in the following assemblies: +// +// * maps/jobs/monitor-model-performance.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-monitor-model-performance_{context}"] += Monitor model performance + +[role="_abstract"] +Export and view model-server metrics so you can identify bottlenecks, optimize resource allocation, and ensure efficient infrastructure utilization. diff --git a/modules/microshift-prepare-and-package-ai-models.adoc b/modules/microshift-prepare-and-package-ai-models.adoc new file mode 100644 index 00000000000..962037151a8 --- /dev/null +++ b/modules/microshift-prepare-and-package-ai-models.adoc @@ -0,0 +1,10 @@ +// Module included in the following assemblies: +// +// * maps/jobs/prepare-and-package-ai-models.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-prepare-and-package-ai-models_{context}"] += Package your AI model for edge deployment + +[role="_abstract"] +Create a namespace and package your AI model into an OCI image in ModelCar format so you can securely store and deploy the model to your {microshift-short} node. diff --git a/modules/microshift-remove-ai-model-serving-components.adoc b/modules/microshift-remove-ai-model-serving-components.adoc new file mode 100644 index 00000000000..29b041cf70b --- /dev/null +++ b/modules/microshift-remove-ai-model-serving-components.adoc @@ -0,0 +1,15 @@ +// Module included in the following assemblies: +// +// * maps/jobs/remove-ai-model-serving-components.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-remove-ai-model-serving-components_{context}"] += Remove AI model serving components and free resources + +[role="_abstract"] +When you no longer need to serve a model, delete the `InferenceService`, clean up OCI images, and uninstall the `microshift-ai-model-serving` RPM to manage resources on constrained edge nodes. + +[NOTE] +==== +Documentation for uninstalling AI model serving components is not yet available in the {microshift-short} AI chapter. +==== diff --git a/modules/microshift-rhoai-create-ns.adoc b/modules/microshift-rhoai-create-ns.adoc index d2d410b471e..f71570bad36 100644 --- a/modules/microshift-rhoai-create-ns.adoc +++ b/modules/microshift-rhoai-create-ns.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-create-namespace_{context}"] -= Creating a namespace for your AI model on {microshift-short} += Create a namespace for your AI model on {microshift-short} [role="_abstract"] Create a namespace for your AI model and all other resources. Namespaces offer resource isolation, resource management, and access control. diff --git a/modules/microshift-rhoai-create-route.adoc b/modules/microshift-rhoai-create-route.adoc index 253f0d43796..fba55a4e5cb 100644 --- a/modules/microshift-rhoai-create-route.adoc +++ b/modules/microshift-rhoai-create-route.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-create-route_{context}"] -= Creating a route to use for AI queries in {microshift-short} += Create a route to use for AI queries in {microshift-short} [role="_abstract"] You can create a route so that your AI model can receive queries and give output by using the `oc expose svc` command or creating a definition in a YAML file and apply it. diff --git a/modules/microshift-rhoai-export-metrics-otel.adoc b/modules/microshift-rhoai-export-metrics-otel.adoc index 96fb5151fcc..b3074704e14 100644 --- a/modules/microshift-rhoai-export-metrics-otel.adoc +++ b/modules/microshift-rhoai-export-metrics-otel.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-export-metrics-otel_{context}"] -= Exporting model-server metrics by using Open Telemetry += Export model-server metrics by using Open Telemetry [role="_abstract"] You can export model-server metrics by using Open Telemetry if you installed the `microshift-observability` RPM for {microshift-short}. diff --git a/modules/microshift-rhoai-get-model-ready-inference.adoc b/modules/microshift-rhoai-get-model-ready-inference.adoc index ca953eed55b..185451048d0 100644 --- a/modules/microshift-rhoai-get-model-ready-inference.adoc +++ b/modules/microshift-rhoai-get-model-ready-inference.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-get-model-ready-inference_{context}"] -= Getting your AI model ready for inference += Get your AI model ready for inference [role="_abstract"] Before querying your AI model through the API, you can get the model ready to provide answers based on the training data. diff --git a/modules/microshift-rhoai-get-model-server-metrics.adoc b/modules/microshift-rhoai-get-model-server-metrics.adoc index b613105e358..4fe7eb86f9d 100644 --- a/modules/microshift-rhoai-get-model-server-metrics.adoc +++ b/modules/microshift-rhoai-get-model-server-metrics.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-get-model-server-metrics_{context}"] -= Getting the model-server metrics += Get the model-server metrics [role="_abstract"] After making a query, you can get the model server's metrics to identify bottlenecks, optimize resource allocation, and ensure efficient infrastructure utilization. diff --git a/modules/microshift-rhoai-inferenceservice-ex.adoc b/modules/microshift-rhoai-inferenceservice-ex.adoc index d62adfcd607..002b9cdb661 100644 --- a/modules/microshift-rhoai-inferenceservice-ex.adoc +++ b/modules/microshift-rhoai-inferenceservice-ex.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-inferenceservice-ex_{context}"] -= Creating an InferenceService custom resource += Create an InferenceService custom resource [role="_abstract"] You can create an `InferenceService` custom resource (CR) that instructs KServe how to create a deployment for serving your AI model. KServe uses the `ServingRuntime` based on the `modelFormat` value specified in the `InferenceService` CR. diff --git a/modules/microshift-rhoai-model-package-oci.adoc b/modules/microshift-rhoai-model-package-oci.adoc index 93629a95be9..c8f0f98d10e 100644 --- a/modules/microshift-rhoai-model-package-oci.adoc +++ b/modules/microshift-rhoai-model-package-oci.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-model-package-oci_{context}"] -= Packaging your AI model into an OCI image += Package your AI model into an OCI image [role="_abstract"] You can package your model into an OCI image and use the ModelCar approach to help you set up offline environments. With the ModelCar approach, your model can be embedded just like any other container image. diff --git a/modules/microshift-rhoai-model-serving-rt-verify.adoc b/modules/microshift-rhoai-model-serving-rt-verify.adoc index 143fcb8c32a..fcfae17080d 100644 --- a/modules/microshift-rhoai-model-serving-rt-verify.adoc +++ b/modules/microshift-rhoai-model-serving-rt-verify.adoc @@ -5,7 +5,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-model-serving-rt-verify_{context}"] -= Verifying that the model-serving runtime is ready += Verify that the model-serving runtime is ready [role="_abstract"] You can use the {oc-first} to verify that your model-serving runtime is ready for use by checking that the downstream generation activities are complete. diff --git a/modules/microshift-rhoai-override-kserve-config.adoc b/modules/microshift-rhoai-override-kserve-config.adoc index 93527d97111..3b4e4b0018a 100644 --- a/modules/microshift-rhoai-override-kserve-config.adoc +++ b/modules/microshift-rhoai-override-kserve-config.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: CONCEPT [id="microshift-rhoai-override-kserve-config_{context}"] -= Overriding KServe configuration += Override KServe configuration [role="_abstract"] You can override KServe settings to customize your model-serving environment. diff --git a/modules/microshift-rhoai-query-model.adoc b/modules/microshift-rhoai-query-model.adoc index 9228494c9e7..0028039b95e 100644 --- a/modules/microshift-rhoai-query-model.adoc +++ b/modules/microshift-rhoai-query-model.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-query-model_{context}"] -= Querying your AI model += Query your AI model [role="_abstract"] You can make an inference request against the AI model server that is using the `ovms-resnet50` model. diff --git a/modules/microshift-rhoai-servingruntimes-ex.adoc b/modules/microshift-rhoai-servingruntimes-ex.adoc index 644b1b0d607..5cb47123ed0 100644 --- a/modules/microshift-rhoai-servingruntimes-ex.adoc +++ b/modules/microshift-rhoai-servingruntimes-ex.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-servingruntimes-ex_{context}"] -= Creating a ServingRuntime CR for use in {microshift-short} += Create a ServingRuntime CR for use in {microshift-short} [role="_abstract"] You can create a `ServingRuntime` custom resource (CR) based on installed manifests and release information. diff --git a/modules/microshift-rhoai-verify-model-connected.adoc b/modules/microshift-rhoai-verify-model-connected.adoc index 90447f9f2fe..5baf4d5b9ce 100644 --- a/modules/microshift-rhoai-verify-model-connected.adoc +++ b/modules/microshift-rhoai-verify-model-connected.adoc @@ -4,7 +4,7 @@ :_mod-docs-content-type: PROCEDURE [id="microshift-rhoai-verify-model-connected_{context}"] -= Verifying that your AI model is accessible += Verify that your AI model is accessible [role="_abstract"] Before querying the model through the API, you can verify that the model is accessible and ready to provide answers based on the connected data. diff --git a/modules/microshift-secure-ai-inference-endpoints.adoc b/modules/microshift-secure-ai-inference-endpoints.adoc new file mode 100644 index 00000000000..55572bf09e1 --- /dev/null +++ b/modules/microshift-secure-ai-inference-endpoints.adoc @@ -0,0 +1,10 @@ +// Module included in the following assemblies: +// +// * maps/jobs/secure-ai-inference-endpoints.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-secure-ai-inference-endpoints_{context}"] += Secure your AI inference endpoints + +[role="_abstract"] +Add TLS encryption and authentication to inference routes so you can protect model access and meet security requirements. diff --git a/modules/microshift-serve-ai-models.adoc b/modules/microshift-serve-ai-models.adoc new file mode 100644 index 00000000000..42169c7d176 --- /dev/null +++ b/modules/microshift-serve-ai-models.adoc @@ -0,0 +1,10 @@ +// Module included in the following assemblies: +// +// * maps/jobs/serve-ai-models.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-serve-ai-models_{context}"] += Serve AI models on {microshift-short} + +[role="_abstract"] +Configure a model-serving runtime and create an `InferenceService` custom resource so you can serve your AI model on {microshift-short}. diff --git a/modules/microshift-update-running-ai-model.adoc b/modules/microshift-update-running-ai-model.adoc new file mode 100644 index 00000000000..d23a82ade6a --- /dev/null +++ b/modules/microshift-update-running-ai-model.adoc @@ -0,0 +1,15 @@ +// Module included in the following assemblies: +// +// * maps/jobs/update-running-ai-model.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-update-running-ai-model_{context}"] += Update a running AI model to a newer version + +[role="_abstract"] +When a newer version of your AI model is available, update the served model without downtime so you can deliver improved inference results. + +[NOTE] +==== +Documentation for updating a running model is not yet available in the {microshift-short} AI chapter. +==== diff --git a/modules/microshift-use-alternative-storage-for-ai-models.adoc b/modules/microshift-use-alternative-storage-for-ai-models.adoc new file mode 100644 index 00000000000..989f8742560 --- /dev/null +++ b/modules/microshift-use-alternative-storage-for-ai-models.adoc @@ -0,0 +1,10 @@ +// Module included in the following assemblies: +// +// * maps/jobs/use-alternative-storage-for-ai-models.adoc + +:_mod-docs-content-type: CONCEPT +[id="microshift-use-alternative-storage-for-ai-models_{context}"] += Use alternative storage for your AI models + +[role="_abstract"] +When you already have S3-compatible storage or a persistent volume claim, you can point your `InferenceService` to that storage instead of packaging a ModelCar OCI image. Only the ModelCar approach is tested and supported for {microshift-short}.