Add alert sideloading, Scaling Improvements (#1924)

assets/src/generated/graphql.ts

@@ -665,6 +665,8 @@ export type BackupAttributes = {
 export type BedrockAiAttributes = {
   /** aws access key id to use, you can also use IRSA for self-hosted consoles */
   accessKeyId?: InputMaybe<Scalars['String']['input']>;
+  /** the model to use for vector embeddings */
+  embeddingModel?: InputMaybe<Scalars['String']['input']>;
   /** the bedrock model id to use */
   modelId: Scalars['String']['input'];
   /** aws secret access key to use, you can also use IRSA for self-hosted consoles */
@@ -4849,6 +4851,8 @@ export type PolicyEngine = {
 };
 
 export type PolicyEngineAttributes = {
+  /** the maximum allowed severity without failing the stack run */
+  maxSeverity?: InputMaybe<VulnSeverity>;
   /** the policy engine to use with this stack */
   type: PolicyEngineType;
 };
@@ -7111,6 +7115,7 @@ export type RootQueryTypeClustersArgs = {
   q?: InputMaybe<Scalars['String']['input']>;
   tag?: InputMaybe<TagInput>;
   tagQuery?: InputMaybe<TagQuery>;
+  upgradeable?: InputMaybe<Scalars['Boolean']['input']>;
 };
 
 
@@ -8964,13 +8969,15 @@ export enum StackType {
 export type StackViolationCause = {
   __typename?: 'StackViolationCause';
   end: Scalars['Int']['output'];
+  filename?: Maybe<Scalars['String']['output']>;
   lines?: Maybe<Array<Maybe<StackViolationCauseLine>>>;
   resource: Scalars['String']['output'];
   start: Scalars['Int']['output'];
 };
 
 export type StackViolationCauseAttributes = {
   end: Scalars['Int']['input'];
+  filename?: InputMaybe<Scalars['String']['input']>;
   lines?: InputMaybe<Array<InputMaybe<StackViolationCauseLineAttributes>>>;
   resource: Scalars['String']['input'];
   start: Scalars['Int']['input'];

charts/controller/crds/deployments.plural.sh_deploymentsettings.yaml

@@ -56,6 +56,9 @@ spec:
                           of the same API scheme (for instance Together.ai uses the
                           OpenAI API spec)
                         type: string
+                      embeddingModel:
+                        description: Model to use for generating embeddings
+                        type: string
                       model:
                         description: Model is the LLM model name to use.
                         type: string
@@ -96,6 +99,9 @@ spec:
                         description: The azure openai Data plane - inference api version
                           to use, defaults to 2024-10-01-preview or the latest available
                         type: string
+                      embeddingModel:
+                        description: Model to use for generating embeddings
+                        type: string
                       endpoint:
                         description: 'Your Azure OpenAI endpoint, should be formatted
                           like: https://{endpoint}/openai/deployments/{deployment-id}"'
@@ -173,6 +179,18 @@ spec:
                     required:
                     - modelId
                     type: object
+                  embeddingProvider:
+                    description: Provider to use for generating embeddings. Oftentimes
+                      foundational model providers do not have embeddings models,
+                      and it's better to simply use OpenAI.
+                    enum:
+                    - OPENAI
+                    - ANTHROPIC
+                    - OLLAMA
+                    - AZURE
+                    - BEDROCK
+                    - VERTEX
+                    type: string
                   enabled:
                     default: false
                     description: Enabled defines whether to enable the AI integration
@@ -227,6 +245,9 @@ spec:
                           of the same API scheme (for instance Together.ai uses the
                           OpenAI API spec)
                         type: string
+                      embeddingModel:
+                        description: Model to use for generating embeddings
+                        type: string
                       model:
                         description: Model is the LLM model name to use.
                         type: string

config/config.exs

@@ -52,6 +52,7 @@ config :console,
   audit_expiry: 30,
   admin_emails: [],
   cloud: false,
+  chunk_size: 1024 * 10,
   sidecar_token: "example",
   sidecar_token_path: "./secrets",
   cache_adapter: Console.Cache,
@@ -112,7 +113,7 @@ config :console, Console.PartitionedCache,
     allocated_memory: 1000 * 1000 * 500
   ]
 
-config :my_app, Console.MultilevelCache,
+config :console, Console.MultilevelCache,
   model: :inclusive,
   levels: [
     {
@@ -161,4 +162,8 @@ config :console, Console.PromEx,
   metrics_server: :disabled,
   version: "0.11.0"
 
+config :console, :ttls,
+  helm: :timer.minutes(30),
+  cluster_metrics: :timer.hours(6)
+
 import_config "#{Mix.env()}.exs"

config/prod.exs

@@ -41,6 +41,7 @@ config :console, Console.Cron.Scheduler,
     {"*/4 * * * *",    {Console.Deployments.Cron, :scan_pipeline_stages, []}},
     {"*/4 * * * *",    {Console.Deployments.Cron, :scan_pending_promotions, []}},
     {"*/4 * * * *",    {Console.Deployments.Cron, :scan_pending_contexts, []}},
+    {"*/5 * * * *",    {Console.Deployments.Cron, :cache_warm, []}},
     {"*/10 * * * *",   {Console.Deployments.Init, :ensure_secret, []}},
     {"*/5 * * * *",    {Console.AI.Cron, :services, []}},
     {"*/5 * * * *",    {Console.AI.Cron, :stacks, []}},

go/controller/api/v1alpha1/deploymentsettings_types.go

@@ -297,6 +297,12 @@ type AISettings struct {
 	// +kubebuilder:validation:Optional
 	ToolProvider *console.AiProvider `json:"toolProvider,omitempty"`
 
+	// Provider to use for generating embeddings. Oftentimes foundational model providers do not have embeddings models, and it's better to simply use OpenAI.
+	//
+	// +kubebuilder:validation:Enum=OPENAI;ANTHROPIC;OLLAMA;AZURE;BEDROCK;VERTEX
+	// +kubebuilder:validation:Optional
+	EmbeddingProvider *console.AiProvider `json:"embeddingProvider,omitempty"`
+
 	// OpenAI holds the OpenAI provider configuration.
 	//
 	// +kubebuilder:validation:Optional
@@ -364,9 +370,10 @@ func (in *LoggingSettings) Attributes(ctx context.Context, c client.Client, name
 
 func (in *AISettings) Attributes(ctx context.Context, c client.Client, namespace string) (*console.AiSettingsAttributes, error) {
 	attr := &console.AiSettingsAttributes{
-		Enabled:      in.Enabled,
-		Provider:     in.Provider,
-		ToolProvider: in.ToolProvider,
+		Enabled:           in.Enabled,
+		Provider:          in.Provider,
+		ToolProvider:      in.ToolProvider,
+		EmbeddingProvider: in.EmbeddingProvider,
 	}
 
 	if in.Tools != nil && in.Tools.CreatePr != nil {
@@ -397,10 +404,11 @@ func (in *AISettings) Attributes(ctx context.Context, c client.Client, namespace
 		}
 
 		attr.Openai = &console.OpenaiSettingsAttributes{
-			AccessToken: &token,
-			Model:       in.OpenAI.Model,
-			BaseURL:     in.OpenAI.BaseUrl,
-			ToolModel:   in.OpenAI.ToolModel,
+			AccessToken:    &token,
+			Model:          in.OpenAI.Model,
+			BaseURL:        in.OpenAI.BaseUrl,
+			ToolModel:      in.OpenAI.ToolModel,
+			EmbeddingModel: in.OpenAI.EmbeddingModel,
 		}
 	case console.AiProviderAnthropic:
 		if in.Anthropic == nil {
@@ -413,9 +421,10 @@ func (in *AISettings) Attributes(ctx context.Context, c client.Client, namespace
 		}
 
 		attr.Anthropic = &console.AnthropicSettingsAttributes{
-			AccessToken: lo.ToPtr(token),
-			Model:       in.Anthropic.Model,
-			ToolModel:   in.Anthropic.ToolModel,
+			AccessToken:    lo.ToPtr(token),
+			Model:          in.Anthropic.Model,
+			ToolModel:      in.Anthropic.ToolModel,
+			EmbeddingModel: in.Anthropic.EmbeddingModel,
 		}
 	case console.AiProviderAzure:
 		if in.Azure == nil {
@@ -428,11 +437,12 @@ func (in *AISettings) Attributes(ctx context.Context, c client.Client, namespace
 		}
 
 		attr.Azure = &console.AzureOpenaiAttributes{
-			Endpoint:    in.Azure.Endpoint,
-			APIVersion:  in.Azure.ApiVersion,
-			Model:       in.Azure.Model,
-			ToolModel:   in.Azure.ToolModel,
-			AccessToken: token,
+			Endpoint:       in.Azure.Endpoint,
+			APIVersion:     in.Azure.ApiVersion,
+			Model:          in.Azure.Model,
+			ToolModel:      in.Azure.ToolModel,
+			EmbeddingModel: in.Azure.EmbeddingModel,
+			AccessToken:    token,
 		}
 	case console.AiProviderVertex:
 		if in.Vertex == nil {
@@ -499,6 +509,11 @@ type AIProviderSettings struct {
 	// +kubebuilder:validation:Optional
 	ToolModel *string `json:"toolModel,omitempty"`
 
+	// Model to use for generating embeddings
+	//
+	// +kubebuilder:validation:Optional
+	EmbeddingModel *string `json:"embeddingModel,omitempty"`
+
 	// A custom base url to use, for reimplementations of the same API scheme (for instance Together.ai uses the OpenAI API spec)
 	//
 	// +kubebuilder:validation:Optional
@@ -556,6 +571,11 @@ type AzureOpenAISettings struct {
 	// +kubebuilder:validation:Optional
 	ToolModel *string `json:"toolModel,omitempty"`
 
+	// Model to use for generating embeddings
+	//
+	// +kubebuilder:validation:Optional
+	EmbeddingModel *string `json:"embeddingModel,omitempty"`
+
 	// TokenSecretRef is a reference to the local secret holding the token to access
 	// the configured AI provider.
 	//

go/controller/config/crd/bases/deployments.plural.sh_deploymentsettings.yaml

@@ -56,6 +56,9 @@ spec:
                           of the same API scheme (for instance Together.ai uses the
                           OpenAI API spec)
                         type: string
+                      embeddingModel:
+                        description: Model to use for generating embeddings
+                        type: string
                       model:
                         description: Model is the LLM model name to use.
                         type: string
@@ -96,6 +99,9 @@ spec:
                         description: The azure openai Data plane - inference api version
                           to use, defaults to 2024-10-01-preview or the latest available
                         type: string
+                      embeddingModel:
+                        description: Model to use for generating embeddings
+                        type: string
                       endpoint:
                         description: 'Your Azure OpenAI endpoint, should be formatted
                           like: https://{endpoint}/openai/deployments/{deployment-id}"'
@@ -173,6 +179,18 @@ spec:
                     required:
                     - modelId
                     type: object
+                  embeddingProvider:
+                    description: Provider to use for generating embeddings. Oftentimes
+                      foundational model providers do not have embeddings models,
+                      and it's better to simply use OpenAI.
+                    enum:
+                    - OPENAI
+                    - ANTHROPIC
+                    - OLLAMA
+                    - AZURE
+                    - BEDROCK
+                    - VERTEX
+                    type: string
                   enabled:
                     default: false
                     description: Enabled defines whether to enable the AI integration
@@ -227,6 +245,9 @@ spec:
                           of the same API scheme (for instance Together.ai uses the
                           OpenAI API spec)
                         type: string
+                      embeddingModel:
+                        description: Model to use for generating embeddings
+                        type: string
                       model:
                         description: Model is the LLM model name to use.
                         type: string

lib/console/ai/provider.ex

@@ -53,12 +53,16 @@ defmodule Console.AI.Provider do
 
   def embeddings(text) do
     settings = Console.Deployments.Settings.cached()
-    with {:ok, %mod{} = client} <- client(settings),
+    with {:ok, %mod{} = client} <- embedding_client(settings),
       do: mod.embeddings(client, text)
   end
 
   def summary(text), do: completion([{:user, text}], preface: @summary)
 
+  defp embedding_client(%DeploymentSettings{ai: %AI{embedding_provider: p}} = settings) when not is_nil(p),
+    do: client(put_in(settings.ai.provider, p))
+  defp embedding_client(settings), do: client(settings)
+
   defp tool_client(%DeploymentSettings{ai: %AI{tool_provider: p}} = settings) when not is_nil(p),
     do: client(put_in(settings.ai.provider, p))
   defp tool_client(settings), do: client(settings)

lib/console/application.ex

@@ -14,13 +14,15 @@ defmodule Console.Application do
       {Phoenix.PubSub, [name: Console.PubSub, adapter: Phoenix.PubSub.PG2]},
       Console.Cache,
       Console.ReplicatedCache,
+      Console.MultilevelCache,
       Console.TestCache,
       Console.LocalCache,
       ConsoleWeb.Endpoint,
       Console.Plural.Config,
       Console.Features,
       Console.Cron.Scheduler,
       Console.Buffers.ClusterAudit,
+      Console.Deployments.Local.Server,
       {Registry, [keys: :unique, name: Console.Buffer.Base.registry()]},
       {Registry, [keys: :unique, name: Console.Deployments.Git.Agent.registry()]},
       {Registry, [keys: :unique, name: Console.Deployments.Pipelines.Supervisor.registry()]},