fix: resolve deployment crashes in coordinator and API services

Coordinator: install all dependencies from pyproject.toml instead of
hardcoded subset (missing slowapi, anthropic, opentelemetry-*).

API: FederationAgentService now gracefully disables when orchestrator
URL is not configured instead of throwing and crashing the app.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

apps/api/src/federation/federation-agent.service.ts

@@ -44,6 +44,7 @@ export interface AgentCommandResponse {
 export class FederationAgentService {
   private readonly logger = new Logger(FederationAgentService.name);
   private readonly orchestratorUrl: string;
+  private readonly enabled: boolean;
 
   constructor(
     private readonly prisma: PrismaService,
@@ -60,18 +61,28 @@ export class FederationAgentService {
     const validationResult = validateUrl(url, isDevelopment);
     if (!validationResult.valid) {
       const errorMessage = validationResult.error ?? "Unknown validation error";
-      this.logger.error(`Invalid orchestrator URL: ${errorMessage}`);
-      // Log security event
+      this.logger.warn(
+        `Federation agent service disabled: ${errorMessage}. Set orchestrator.url to enable.`
+      );
       this.auditService.logInvalidOrchestratorUrl(url, errorMessage);
-      throw new Error(errorMessage);
+      this.orchestratorUrl = "";
+      this.enabled = false;
+      return;
     }
 
     this.orchestratorUrl = url;
+    this.enabled = true;
     this.logger.log(
       `FederationAgentService initialized with orchestrator URL: ${this.orchestratorUrl}`
     );
   }
 
+  private assertEnabled(): void {
+    if (!this.enabled) {
+      throw new Error("Federation agent service is disabled: orchestrator URL not configured");
+    }
+  }
+
   /**
    * Spawn an agent on a remote federated instance
    * @param workspaceId Workspace ID
@@ -84,6 +95,7 @@ export class FederationAgentService {
     connectionId: string,
     payload: SpawnAgentCommandPayload
   ): Promise<CommandMessageDetails> {
+    this.assertEnabled();
     this.logger.log(
       `Spawning agent on remote instance via connection ${connectionId} for task ${payload.taskId}`
     );
@@ -126,6 +138,7 @@ export class FederationAgentService {
     connectionId: string,
     agentId: string
   ): Promise<CommandMessageDetails> {
+    this.assertEnabled();
     this.logger.log(`Getting agent status for ${agentId} via connection ${connectionId}`);
 
     // Validate connection exists and is active
@@ -167,6 +180,7 @@ export class FederationAgentService {
     connectionId: string,
     agentId: string
   ): Promise<CommandMessageDetails> {
+    this.assertEnabled();
     this.logger.log(`Killing agent ${agentId} via connection ${connectionId}`);
 
     // Validate connection exists and is active
@@ -208,6 +222,7 @@ export class FederationAgentService {
     commandType: string,
     payload: Record<string, unknown>
   ): Promise<AgentCommandResponse> {
+    this.assertEnabled();
     this.logger.log(`Handling agent command ${commandType} from ${remoteInstanceId}`);
 
     // Verify connection exists for remote instance

apps/coordinator/Dockerfile

@@ -15,14 +15,9 @@ COPY pyproject.toml .
 # Create virtual environment and install dependencies
 RUN python -m venv /opt/venv
 ENV PATH="/opt/venv/bin:$PATH"
+COPY src/ ./src/
 RUN pip install --no-cache-dir --upgrade pip && \
-    pip install --no-cache-dir hatchling && \
-    pip install --no-cache-dir \
-    fastapi>=0.109.0 \
-    uvicorn[standard]>=0.27.0 \
-    pydantic>=2.5.0 \
-    pydantic-settings>=2.1.0 \
-    python-dotenv>=1.0.0
+    pip install --no-cache-dir .
 
 # Production stage
 FROM python:3.11-slim