fix(deploy-progress): decouple deploys from the live progress stream (#1246)

* fix(deploy-progress): decouple deploys from the live progress stream

The deploy progress modal streamed compose output over a WebSocket, but
the deploy itself was coupled to that socket in two ways that could break
or silently abort a deploy:

- The deploy request was gated on the progress socket connecting, so any
  upgrade failure (a reverse proxy blocking WebSocket upgrades, or the
  admin-only stream rejecting a scoped deployer) left the modal stuck on
  "Connecting..." and the deploy never fired.
- The backend terminated the running compose process when that socket
  closed, so minimizing the modal, navigating away, or a network blip
  aborted an in-flight deploy.

Make the progress socket output-only: the deploy is owned by its request
and runs to completion (or the existing command timeout) regardless of
the stream. The modal now degrades to a "Live progress unavailable" state
and still reports success or failure from the request result. Connect
failures, drops, and a connect timeout all release the deploy instead of
blocking it.

Also route progress output per deploy: the frontend sends a correlation
id on both the connectTerminal message and the deploy request header, and
the backend keys progress sockets by that id so concurrent deploys from
different tabs or users no longer cross-stream each other's output.

Cap the in-memory parsed log rows so a very long deploy cannot grow the
modal's state unbounded.

* fix(deploy-progress): generate the deploy session id with a CSPRNG

The per-deploy correlation id keys which WebSocket receives a deploy's
live output, so a guessable id lets one authenticated client register a
victim's id and read its compose output. It was built from Math.random()
plus a timestamp, which is not cryptographically secure.

Generate it with crypto.getRandomValues (128 bits, hex). That is the one
Crypto member available in insecure contexts, so it still works over LAN
HTTP where crypto.randomUUID is unavailable.

* fix(deploy-progress): stop headerless ops bleeding into a keyed progress modal

Address review findings on the progress-stream routing:

- Only an id-less connectTerminal registration may become the id-less
  fallback socket. Previously every connectTerminal (including keyed deploy
  modals) set the fallback, so a headerless operation (bulk update, rollback,
  or a legacy client) resolved via getTerminalWs() into another user's keyed
  deploy modal. Keyed sockets are now excluded from the fallback, and a socket
  that adopts a session id is removed from it.
- The connect-timeout fallback now also flags the modal as "Live progress
  unavailable" instead of leaving it on "Connecting..." while the deploy runs.
- Log only a short prefix of the deploy session id in developer diagnostics,
  not the full capability value.
This commit is contained in:
Anso
2026-05-28 20:51:45 -04:00
committed by GitHub
parent b034de58f3
commit 5dea040ec8
19 changed files with 577 additions and 89 deletions
@@ -254,22 +254,21 @@ describe('ComposeService - runCommand', () => {
await expectation;
});
it('kills and rejects running commands when the WebSocket disconnects', async () => {
it('keeps the command running when the WebSocket disconnects (progress socket is output-only)', async () => {
const proc = createMockProcess();
mockSpawn.mockReturnValue(proc);
const ws = createMockWs();
const svc = ComposeService.getInstance(1);
const promise = svc.runCommand('my-stack', 'restart', ws);
const expectation = expect(promise).rejects.toThrow('client disconnected');
let settled = false;
promise.finally(() => { settled = true; }).catch(() => undefined);
// The deploy is owned by its HTTP request; closing the progress socket
// (panel minimized, navigated away, connection blip) must not abort it.
ws.emit('close');
expect(proc.kill).not.toHaveBeenCalled();
expect(proc.kill).toHaveBeenCalledWith('SIGTERM');
expect(settled).toBe(false);
proc.emit('close', null);
await expectation;
// The command still completes on its own exit, not the socket close.
proc.emit('close', 0);
await expect(promise).resolves.toBeUndefined();
});
it('rewrites ENOMEM spawn failures as host out-of-memory', async () => {
@@ -0,0 +1,84 @@
import { describe, it, expect, beforeEach } from 'vitest';
import { EventEmitter } from 'events';
import WebSocket, { type WebSocketServer } from 'ws';
import { attachGenericConnectionHandlers, getTerminalWs } from '../websocket/generic';
class FakeWs extends EventEmitter {
readyState: number = WebSocket.OPEN;
}
/** Simulate a client opening the generic /ws socket and registering for deploy
* output via the connectTerminal handshake. */
function connect(wss: EventEmitter, sessionId?: string): FakeWs {
const ws = new FakeWs();
wss.emit('connection', ws);
ws.emit('message', Buffer.from(JSON.stringify({ action: 'connectTerminal', ...(sessionId ? { sessionId } : {}) })));
return ws;
}
describe('generic ws terminal registry', () => {
let wss: EventEmitter;
beforeEach(() => {
wss = new EventEmitter();
attachGenericConnectionHandlers(wss as unknown as WebSocketServer);
});
it('routes output to the socket matching the deploy session id', () => {
const a = connect(wss, 'sess-a');
const b = connect(wss, 'sess-b');
expect(getTerminalWs('sess-a')).toBe(a);
expect(getTerminalWs('sess-b')).toBe(b);
expect(getTerminalWs('not-a-session')).toBeUndefined();
});
it('falls back to the most recent id-less socket', () => {
const a = connect(wss);
expect(getTerminalWs()).toBe(a);
const b = connect(wss);
expect(getTerminalWs()).toBe(b);
});
it('drops a session from the registry when its socket closes', () => {
const a = connect(wss, 'sess-x');
expect(getTerminalWs('sess-x')).toBe(a);
a.emit('close');
expect(getTerminalWs('sess-x')).toBeUndefined();
});
it('ignores a registered socket that is no longer open', () => {
const a = connect(wss, 'sess-y');
a.readyState = WebSocket.CLOSED;
expect(getTerminalWs('sess-y')).toBeUndefined();
});
it('rebinds a socket to a new session id and drops the old mapping', () => {
const a = connect(wss, 'sess-1');
// Same socket re-registers under a new id (a second deploy in the same tab).
a.emit('message', Buffer.from(JSON.stringify({ action: 'connectTerminal', sessionId: 'sess-2' })));
expect(getTerminalWs('sess-1')).toBeUndefined();
expect(getTerminalWs('sess-2')).toBe(a);
});
it('clears the id-less fallback when its socket closes', () => {
const a = connect(wss);
expect(getTerminalWs()).toBe(a);
a.emit('close');
expect(getTerminalWs()).toBeUndefined();
});
it('does not expose a keyed socket as the id-less fallback', () => {
// A headerless operation (bulk / rollback / legacy) resolves via getTerminalWs()
// with no id; it must never reach a keyed deploy modal's socket.
const a = connect(wss, 'sess-keyed');
expect(getTerminalWs('sess-keyed')).toBe(a);
expect(getTerminalWs()).toBeUndefined();
});
it('drops the id-less fallback when that socket later adopts a session id', () => {
const a = connect(wss);
expect(getTerminalWs()).toBe(a);
a.emit('message', Buffer.from(JSON.stringify({ action: 'connectTerminal', sessionId: 'sess-late' })));
expect(getTerminalWs()).toBeUndefined();
expect(getTerminalWs('sess-late')).toBe(a);
});
});