michal/mcpctl
1
0
Fork 0
Code Issues Pull Requests 7 Actions Packages Projects Releases Wiki Activity

feat: eager vLLM warmup and smart page titles in paginate stage

Browse Source 
- Add warmup() to LlmProvider interface for eager subprocess startup
- ManagedVllmProvider.warmup() starts vLLM in background on project load
- ProviderRegistry.warmupAll() triggers all managed providers
- NamedProvider proxies warmup() to inner provider
- paginate stage generates LLM-powered descriptive page titles when
  available, cached by content hash, falls back to generic "Page N"
- project-mcp-endpoint calls warmupAll() on router creation so vLLM
  is loading while the session initializes

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Michal
2026-03-03 19:07:39 +00:00
parent 0427d7dc1a
commit 03827f11e4
147 changed files with 17561 additions and 2093 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
src/mcpd/src/services/instance.service.ts
View File

@@ -4,8 +4,11 @@ import type { McpOrchestrator, ContainerSpec, ContainerInfo } from './orchestrat
import { NotFoundError } from './mcp-server.service.js';
import { resolveServerEnv } from './env-resolver.js';
/** Default image for npm-based MCP servers (STDIO with packageName, no dockerImage). */
const DEFAULT_NODE_RUNNER_IMAGE = process.env['MCPD_NODE_RUNNER_IMAGE'] ?? 'mysources.co.uk/michal/mcpctl-node-runner:latest';
/** Runner images for package-based MCP servers, keyed by runtime name. */
const RUNNER_IMAGES: Record<string, string> = {
node: process.env['MCPD_NODE_RUNNER_IMAGE'] ?? 'mysources.co.uk/michal/mcpctl-node-runner:latest',
python: process.env['MCPD_PYTHON_RUNNER_IMAGE'] ?? 'mysources.co.uk/michal/mcpctl-python-runner:latest',
};
/** Network for MCP server containers (matches docker-compose mcp-servers network). */
const MCP_SERVERS_NETWORK = process.env['MCPD_MCP_NETWORK'] ?? 'mcp-servers';
@@ -183,18 +186,19 @@ export class InstanceService {
// Determine image + command based on server config:
// 1. Explicit dockerImage → use as-is
// 2. packageName (npm) → use node-runner image + npx command
// 2. packageName → use runtime-specific runner image (node/python/go/...)
// 3. Fallback → server name (legacy)
let image: string;
let npmCommand: string[] | undefined;
let pkgCommand: string[] | undefined;
if (server.dockerImage) {
image = server.dockerImage;
} else if (server.packageName) {
image = DEFAULT_NODE_RUNNER_IMAGE;
// Build npx command: entrypoint is ["npx", "-y"], so CMD = [packageName, ...args]
const runtime = (server.runtime as string | null) ?? 'node';
image = RUNNER_IMAGES[runtime] ?? RUNNER_IMAGES['node']!;
// Runner entrypoint handles package execution (npx -y / uvx / go run)
const serverCommand = server.command as string[] | null;
npmCommand = [server.packageName, ...(serverCommand ?? [])];
pkgCommand = [server.packageName, ...(serverCommand ?? [])];
} else {
image = server.name;
}
@@ -218,10 +222,10 @@ export class InstanceService {
if (server.transport === 'SSE' || server.transport === 'STREAMABLE_HTTP') {
spec.containerPort = server.containerPort ?? 3000;
}
// npm-based servers: command = [packageName, ...args] (entrypoint handles npx -y)
// Package-based servers: command = [packageName, ...args] (entrypoint handles execution)
// Docker-image servers: use explicit command if provided
if (npmCommand) {
spec.command = npmCommand;
if (pkgCommand) {
spec.command = pkgCommand;
} else {
const command = server.command as string[] | null;
if (command) {