Files
AgentHub/server/__tests__/subagent-cost-backfill.test.js
T
nntrivi2001 ca821008e2 perf: v2.5.10 — ready for 10+ concurrent users, leaner locally
Measured on a copy of a real 1.2 GB / 253K-event database, 10 simulated
browsers + 10 hooks/s for 90 s: event-loop delay p99 1239 ms → 32 ms (max
2.0 s → 94 ms), 266 failed requests → 0, 74 → 111 requests/s served, server
RSS 633 → 346 MB. Entry bundle 975 KB raw → 394 KB (106 KB brotli).

Server
- /api/analytics: the events-table scans run in the read worker
  (scoped-stats eventAggregates, read-pool job analyticsEvents); runRead takes
  maxAgeMs — analytics 10 s, /api/workflows 30 s — keyed by the args
- /api/workflows/skill-usage: +event_type and a tool_name range instead of
  LIKE, so it uses idx_events_tool (784 ms → a few ms)
- idx_events_session_type_tool (session_id, event_type, tool_name,
  created_at) replaces idx_events_session_type: covering reads for session
  stats, the hook dedupes and the import resume point (1.2 s → 20 ms); 64 MB
  cache_size and 256 MB mmap_size on both connections
- WAL: the server sets wal_autocheckpoint = 0 and checkpoints PASSIVE every
  DASHBOARD_WAL_CHECKPOINT_MS (5 s) from the read worker's own writable
  connection; a failed checkpoint restores the automatic one (hook COMMIT p95
  109 ms → 15 ms at 100 hooks/s)
- tmux: one list-sessions answers a whole GET /api/lanes and the heal beat
  (liveSessions), shared for 1 s and cleared by this process's new/kill —
  GET /api/run's sync spawns had blocked 60 s of a 90 s run; the heal beat
  skips tmux when nothing runs; attach options chained into 2 spawns
- hooks: TurnDuration/APIError deduped against one read per hook instead of a
  query per transcript entry (23 ms/hook on a 763-turn session); one lane
  scan per hook; hourly upserts skip rows that did not grow; frames are sent
  after the transaction commits (and after the pane binding); new_event
  carries its row id; a split pane's (slot 1–9) hooks no longer rebind the
  lane's session_id
- watchdog liveness probe runs ps asynchronously; transcript newline scan via
  indexOf; startup subagent backfill fingerprints the files it already checked
- pty attach: backpressure (pause past 1 MB unsent, resume under 256 KB)
- static: hashed assets served from their pre-compressed .br/.gz sibling

Client
- hidden tabs neither refetch on WS events nor poll (useLiveRefresh, panes,
  lane cards, git header, changes panel, settings); no per-tab hourly
  /updates/check (the server already checks)
- refetch storms throttled (workspace tree/session list 3 s, /api/run only on
  a run-key change, notification pane lookup only when shown); WS reconnect
  jitter; terminal retry backoff 1.5 → 30 s, stops on 1008
- heavier throttles for aggregates (15 s) and detail views; Monitor keepIfSame;
  GET in-flight dedupe in request()
- lazy Dashboard/History, xterm (with the first terminal), lane modals, Sagi
  (only while enabled); only the active locale loads; Workspace chunk
  prefetched on a /workspace deep link
- Sagi: no work while disabled, pupil via CSS variables, memo avatar
- build: .br/.gz siblings, sourcemap hidden; service worker no longer caches
  /assets/ (cache v3 drops the old one)

Tests: scripts/test-server.js drops the GIT_DIR-family variables a hook
exports, so a suite run from a linked worktree's pre-commit can no longer
commit into the real repository or set core.bare.

Docs: ARCHITECTURE, DATABASE, API, DEPLOYMENT, client README,
docs/plans/perf/README.md (audit, checklist, results).
2026-10-07 13:24:15 +07:00

169 lines
6.2 KiB
JavaScript

/**
* @file Tests the one-time backfill that stamps metadata.tokens onto subagent
* rows predating per-agent cost tracking.
*
* A historical session whose transcript never changes again is mtime-skipped by
* the continuous sync, so its subagents (imported before per-agent cost existed)
* would never gain a tokens bucket and their cards would show no cost. The
* startup backfill re-parses those transcripts and stamps the metadata — WITHOUT
* touching session token_usage. This suite verifies:
*
* 1. A pre-feature subagent row (no tokens key) gets its metadata.tokens
* stamped from its transcript, so attachAgentCosts can price it.
* 2. The backfill is metadata-only: it does not create/alter token_usage rows.
* 3. It is self-limiting — a second run stamps nothing new.
*
* @author Nguyễn Ngọc Trí Vĩ <vinnt@smartgift.vn>
*/
const { describe, it, before, after } = require("node:test");
const assert = require("node:assert/strict");
const path = require("path");
const fs = require("fs");
const os = require("os");
const TEST_DB = path.join(os.tmpdir(), `dashboard-subcost-bf-${Date.now()}-${process.pid}.db`);
process.env.DASHBOARD_DB_PATH = TEST_DB;
const dbModule = require("../db");
const { db, stmts } = dbModule;
const importHistory = require("../../scripts/import-history");
const { attachAgentCosts } = require("../routes/pricing");
const SESSION = "dddddddd-4444-4444-8444-dddddddddddd";
const MAIN = `${SESSION}-main`;
const SUB_ID = "aaaa1111-bbbb-2222-cccc-333344445555";
const SUB_MODEL = "claude-haiku-4-5-20251001";
let tmpDir;
let transcriptPath;
after(() => {
if (db) db.close();
for (const suffix of ["", "-wal", "-shm"]) {
try {
fs.unlinkSync(TEST_DB + suffix);
} catch {
/* ignore */
}
}
});
before(() => {
tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), "subcost-bf-"));
// Lay out the on-disk transcript tree the backfill walks:
// <proj>/<session>.jsonl (main transcript path)
// <proj>/<session>/subagents/agent-<id>.jsonl
transcriptPath = path.join(tmpDir, `${SESSION}.jsonl`);
fs.writeFileSync(transcriptPath, "");
const subDir = path.join(tmpDir, SESSION, "subagents");
fs.mkdirSync(subDir, { recursive: true });
fs.writeFileSync(
path.join(subDir, `agent-${SUB_ID}.jsonl`),
[
{
type: "user",
sessionId: SESSION,
timestamp: "2026-04-18T12:00:00.000Z",
message: { content: "go" },
},
{
type: "assistant",
sessionId: SESSION,
timestamp: "2026-04-18T12:00:05.000Z",
message: {
model: SUB_MODEL,
content: [{ type: "text", text: "ok" }],
usage: { input_tokens: 2_000_000, output_tokens: 1_000_000 },
},
},
]
.map((o) => JSON.stringify(o))
.join("\n")
);
// Session with transcript_path set, and a PRE-FEATURE subagent row: metadata
// has a model but NO tokens key (exactly what old imports produced).
stmts.insertSession.run(
SESSION,
"Backfill session",
"completed",
"/tmp/x",
"claude-opus-4-8",
null
);
db.prepare("UPDATE sessions SET transcript_path = ? WHERE id = ?").run(transcriptPath, SESSION);
stmts.insertAgent.run(MAIN, SESSION, "Main Agent", "main", null, "completed", null, null, null);
stmts.insertAgent.run(
`${SESSION}-jsonl-${SUB_ID}`,
SESSION,
"general-purpose",
"subagent",
"general-purpose",
"completed",
"recon",
MAIN,
JSON.stringify({ imported: true, source: "jsonl", model: SUB_MODEL })
);
// Deterministic Haiku pricing: $1 in / $5 out per MTok.
stmts.upsertPricing.run("claude-haiku%", "Claude Haiku 4.5", 1, 5, 0.1, 1.25, 2, 0, 0);
});
describe("subagent token backfill", () => {
it("stamps metadata.tokens on a pre-feature subagent row from its transcript", async () => {
const before = stmts.getAgent.get(`${SESSION}-jsonl-${SUB_ID}`);
assert.ok(!/"tokens":/.test(before.metadata), "row starts without a tokens key");
const res = await importHistory.backfillSubagentTokenMetadata(dbModule);
assert.ok(res.stamped >= 1, "at least one subagent re-parsed");
const row = stmts.getAgent.get(`${SESSION}-jsonl-${SUB_ID}`);
const meta = JSON.parse(row.metadata);
assert.ok(Array.isArray(meta.tokens) && meta.tokens.length === 1, "tokens stamped");
assert.equal(meta.tokens[0].input_tokens, 2_000_000);
assert.equal(meta.tokens[0].output_tokens, 1_000_000);
// Now priceable: 2M in @ $1 + 1M out @ $5 = $2 + $5 = $7.
const withCosts = attachAgentCosts(stmts.listAgentsBySession.all(SESSION));
const sub = withCosts.find((a) => a.id === `${SESSION}-jsonl-${SUB_ID}`);
assert.equal(sub.cost, 7);
});
it("is metadata-only — it creates no token_usage rows", () => {
const rows = db
.prepare("SELECT COUNT(*) AS n FROM token_usage WHERE session_id = ?")
.get(SESSION);
assert.equal(rows.n, 0, "session token_usage untouched by the backfill");
});
it("is self-limiting — a second run stamps nothing new", async () => {
const res = await importHistory.backfillSubagentTokenMetadata(dbModule);
// The one session no longer matches the driving query (its subagent now has
// a tokens key), so no session is re-scanned.
assert.equal(res.sessions, 0);
});
it("parses a row it can never stamp once, until its files change", async () => {
// A subagent row whose transcript holds no usage stays tokens-less forever,
// which used to re-parse its whole session tree on every boot.
stmts.insertAgent.run(
`${SESSION}-jsonl-ghost`,
SESSION,
"ghost",
"subagent",
"general-purpose",
"completed",
null,
MAIN,
JSON.stringify({ imported: true, source: "jsonl" })
);
fs.writeFileSync(path.join(tmpDir, SESSION, "subagents", "agent-ghost.jsonl"), "");
assert.equal((await importHistory.backfillSubagentTokenMetadata(dbModule)).sessions, 1);
assert.equal((await importHistory.backfillSubagentTokenMetadata(dbModule)).sessions, 0);
// A new subagent file appears: the session is looked at again.
fs.writeFileSync(path.join(tmpDir, SESSION, "subagents", "agent-new.jsonl"), "");
assert.equal((await importHistory.backfillSubagentTokenMetadata(dbModule)).sessions, 1);
});
});