feat: analytics parity — tool cost, date ranges, budgets, query explorer, replay, snapshots (v2.4.0)

Analytics parity phases 2–7 (docs/plans/analytics-parity/), built on 138e706 and ported onto v2.3.1: - Cost by AI tool (?tool= on cost/analytics/stats/workflows, by_tool, Codex fast/priority and flex tiers, agenthub_cost_usd_total{tool,model}) - Date ranges (from/to) on cost/analytics/workflows; Analytics range select with Today / This week (Monday) / This month (1st) / 7d / 30d / 90d / All / Custom; workflow filters carried into workflow-stats.js (read worker) - Budgets and cost alerts as alert rule types (alert_rules rebuilt without its rule_type CHECK), GET /api/budgets/status, rule preview, spend pace, status-bar budget chip; rules stay editable in place - Query Explorer (/query, allowlisted sort, CSV/JSON export) and scheduled reports (report_definitions/report_runs, Settings → Reports) - Session Replay tab and immutable snapshots (session_snapshots, /snapshot/:token) - Remote Codex sync, push ingest with a hashed token, privacy redaction and ignored folders (Settings → Data), Cursor sessions (no cost) - Read-only MCP server, statusline, Grafana/Prometheus files Also: resizable sidebar, pet run/wave animations, the workspace home keeps a half-typed prompt, and the pre-commit hook runs only the related tests for a small commit.
2026-10-06 17:12:28 +07:00
parent 837da6e5de
commit 4aae61297f
146 changed files with 13565 additions and 537 deletions
@@ -159,6 +159,10 @@ db.function(
  { deterministic: true },
  require("./lib/event-payload").trimEventData
 );
+// ...and its summary through `redact_text(?)`, which is the identity unless the
+// user turned on secret redaction (`server/lib/privacy.js`; the payload's own
+// redaction is chained inside `trim_event_data`).
+db.function("redact_text", { deterministic: true }, require("./lib/privacy").redactSummary);

 db.exec(`
  CREATE TABLE IF NOT EXISTS sessions (
@@ -344,12 +348,13 @@ db.exec(`
  CREATE INDEX IF NOT EXISTS idx_dashboard_runs_session ON dashboard_runs(session_id);

  -- Rules-based alerting engine. Rules are evaluated server-side: event-driven
-  -- types (event_pattern, token_threshold) on hook ingest, time-based types
-  -- (inactivity, status_duration) on a periodic sweep in server/lib/alerts.js.
+  -- types (event_pattern, token_threshold, cost_threshold) on hook ingest,
+  -- time-based types (inactivity, status_duration, budget) on a periodic
+  -- sweep in server/lib/alerts.js.
  CREATE TABLE IF NOT EXISTS alert_rules (
    id TEXT PRIMARY KEY,
    name TEXT NOT NULL,
-    rule_type TEXT NOT NULL CHECK(rule_type IN ('event_pattern','inactivity','status_duration','token_threshold')),
+    rule_type TEXT NOT NULL, -- validated by validateRuleConfig (server/lib/alerts.js)
    config TEXT NOT NULL DEFAULT '{}',
    enabled INTEGER NOT NULL DEFAULT 1,
    cooldown_seconds INTEGER NOT NULL DEFAULT 300,
@@ -431,6 +436,55 @@ db.exec(`
  CREATE INDEX IF NOT EXISTS idx_webhook_deliveries_target ON webhook_deliveries(target_id, created_at DESC);
  CREATE INDEX IF NOT EXISTS idx_webhook_deliveries_created ON webhook_deliveries(created_at DESC);

+  -- Frozen copies of a session (server/routes/snapshots.js): {session, agents,
+  -- events, stats, cost} as one JSON payload, read back by an unguessable
+  -- token. Immutable once written (trigger below). session_id has no FK on
+  -- purpose: a snapshot outlives its session and survives Clear Data, like
+  -- alert_rules.
+  CREATE TABLE IF NOT EXISTS session_snapshots (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    token TEXT NOT NULL UNIQUE,
+    session_id TEXT NOT NULL,
+    created_at TEXT NOT NULL DEFAULT (strftime('%Y-%m-%dT%H:%M:%fZ', 'now')),
+    payload TEXT NOT NULL
+  );
+  CREATE TRIGGER IF NOT EXISTS session_snapshots_immutable
+    BEFORE UPDATE ON session_snapshots
+    BEGIN SELECT RAISE(ABORT, 'session snapshots are immutable'); END;
+
+  -- Scheduled reports (server/lib/reports.js): a definition says what window
+  -- of days to summarise and when (local hour, and weekday for weekly, at the
+  -- browser's tz_offset); last_slot is the scheduled instant last claimed,
+  -- so two dashboards on one DB never both run it. Each run is logged in
+  -- report_runs; its files live under <data dir>/reports/<report id>/<stamp>.
+  -- Both survive Clear Data, like alert_rules and session_snapshots.
+  CREATE TABLE IF NOT EXISTS report_definitions (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    name TEXT NOT NULL,
+    window_days INTEGER NOT NULL DEFAULT 7,
+    frequency TEXT NOT NULL DEFAULT 'weekly' CHECK(frequency IN ('daily','weekly')),
+    hour INTEGER NOT NULL DEFAULT 8,
+    weekday INTEGER NOT NULL DEFAULT 1,
+    tz_offset INTEGER NOT NULL DEFAULT 0,
+    formats TEXT NOT NULL DEFAULT '["html"]',
+    enabled INTEGER NOT NULL DEFAULT 1,
+    last_slot TEXT,
+    created_at TEXT NOT NULL DEFAULT (strftime('%Y-%m-%dT%H:%M:%fZ', 'now'))
+  );
+  CREATE TABLE IF NOT EXISTS report_runs (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    report_id INTEGER NOT NULL REFERENCES report_definitions(id) ON DELETE CASCADE,
+    created_at TEXT NOT NULL DEFAULT (strftime('%Y-%m-%dT%H:%M:%fZ', 'now')),
+    trigger TEXT NOT NULL,
+    status TEXT NOT NULL,
+    window_from TEXT,
+    window_to TEXT,
+    stamp TEXT,
+    files TEXT NOT NULL DEFAULT '[]',
+    error TEXT
+  );
+  CREATE INDEX IF NOT EXISTS idx_report_runs_report ON report_runs(report_id, id);
+
  -- Workflow-tool runs: fleets of sub-agents spawned by the Claude Code
  -- "Workflow" tool (and self-paced /loop). These emit NO hooks; the source of
  -- truth is the on-disk run journal (~/.claude/projects/<enc-cwd>/<sessionId>/
@@ -1267,6 +1321,17 @@ db.exec(`
  );
 `);

+// Which Cursor agent transcripts have been read, at what mtime
+// (server/lib/cursor-usage.js) — the same "deleted stays deleted" rule.
+db.exec(`
+  CREATE TABLE IF NOT EXISTS cursor_transcripts (
+    path TEXT PRIMARY KEY,
+    session_id TEXT NOT NULL,
+    mtime_ms INTEGER NOT NULL DEFAULT 0,
+    updated_at TEXT NOT NULL DEFAULT (strftime('%Y-%m-%dT%H:%M:%fZ', 'now'))
+  );
+`);
+
 // Remote data sources: other machines whose Claude Code history this dashboard
 // pulls in over SSH. Config only — NO secrets are stored here: authentication
 // always defers to the host's own SSH stack (~/.ssh/config, ssh-agent, keys,
@@ -1290,6 +1355,19 @@ db.exec(`
    updated_at TEXT NOT NULL DEFAULT (strftime('%Y-%m-%dT%H:%M:%fZ', 'now'))
  );
 `);
+// Migrate: a source may also pull Codex history (`sync_codex`, opt-in) from
+// `remote_codex_home` (NULL = ~/.codex). Additive; existing sources stay
+// Claude-only.
+try {
+  db.prepare("SELECT sync_codex FROM remote_sources LIMIT 1").get();
+} catch {
+  db.prepare("ALTER TABLE remote_sources ADD COLUMN sync_codex INTEGER NOT NULL DEFAULT 0").run();
+}
+try {
+  db.prepare("SELECT remote_codex_home FROM remote_sources LIMIT 1").get();
+} catch {
+  db.prepare("ALTER TABLE remote_sources ADD COLUMN remote_codex_home TEXT").run();
+}

 // Migrate webhook_targets for first-class providers. Earlier installs created
 // the table with a 4-value `type` CHECK (slack/discord/teams/generic) and no
@@ -1330,6 +1408,38 @@ db.exec(`
  }
 }

+// Migrate alert_rules off its 4-type `rule_type` CHECK (cost_threshold and
+// budget came later; `validateRuleConfig` is the one list now). Rebuilt with
+// foreign keys off, so alert_events — which references it with ON DELETE
+// CASCADE — keeps every row through the DROP.
+{
+  const meta = db
+    .prepare("SELECT sql FROM sqlite_master WHERE type='table' AND name='alert_rules'")
+    .get();
+  if (meta?.sql?.includes("CHECK(rule_type IN")) {
+    db.exec(`
+      PRAGMA foreign_keys = OFF;
+      BEGIN;
+      CREATE TABLE alert_rules_new (
+        id TEXT PRIMARY KEY,
+        name TEXT NOT NULL,
+        rule_type TEXT NOT NULL,
+        config TEXT NOT NULL DEFAULT '{}',
+        enabled INTEGER NOT NULL DEFAULT 1,
+        cooldown_seconds INTEGER NOT NULL DEFAULT 300,
+        created_at TEXT NOT NULL DEFAULT (strftime('%Y-%m-%dT%H:%M:%fZ', 'now')),
+        updated_at TEXT NOT NULL DEFAULT (strftime('%Y-%m-%dT%H:%M:%fZ', 'now'))
+      );
+      INSERT INTO alert_rules_new (id, name, rule_type, config, enabled, cooldown_seconds, created_at, updated_at)
+        SELECT id, name, rule_type, config, enabled, cooldown_seconds, created_at, updated_at FROM alert_rules;
+      DROP TABLE alert_rules;
+      ALTER TABLE alert_rules_new RENAME TO alert_rules;
+      COMMIT;
+      PRAGMA foreign_keys = ON;
+    `);
+  }
+}
+
 // Migrate: replace legacy idle/connected agent statuses with waiting/working
 // and update the CHECK constraint to the 4-status model.
 // SQLite doesn't support ALTER CHECK, so we detect the old constraint and
@@ -1584,8 +1694,8 @@ const stmts = {
  ),
  getRemoteSource: db.prepare("SELECT * FROM remote_sources WHERE id = ?"),
  insertRemoteSource: db.prepare(
-    `INSERT INTO remote_sources (id, label, host, ssh_port, identity_file, remote_home, enabled)
-     VALUES (?, ?, ?, ?, ?, ?, ?)`
+    `INSERT INTO remote_sources (id, label, host, ssh_port, identity_file, remote_home, enabled, sync_codex, remote_codex_home)
+     VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)`
  ),
  updateRemoteSource: db.prepare(
    `UPDATE remote_sources SET
@@ -1595,6 +1705,8 @@ const stmts = {
       identity_file = ?,
       remote_home = ?,
       enabled = COALESCE(?, enabled),
+       sync_codex = COALESCE(?, sync_codex),
+       remote_codex_home = ?,
       updated_at = strftime('%Y-%m-%dT%H:%M:%fZ', 'now')
     WHERE id = ?`
  ),
@@ -1702,7 +1814,7 @@ const stmts = {
       AND created_at >= strftime('%Y-%m-%dT%H:%M:%fZ', 'now', '-10 seconds')`
  ),
  insertEvent: db.prepare(
-    "INSERT INTO events (session_id, agent_id, event_type, tool_name, summary, data, created_at) VALUES (?, ?, ?, ?, ?, trim_event_data(?), strftime('%Y-%m-%dT%H:%M:%fZ', 'now'))"
+    "INSERT INTO events (session_id, agent_id, event_type, tool_name, summary, data, created_at) VALUES (?, ?, ?, ?, redact_text(?), trim_event_data(?), strftime('%Y-%m-%dT%H:%M:%fZ', 'now'))"
  ),
  // Same insert with an EXPLICIT created_at, for events recovered from a
  // transcript: their real time is in the file, not the moment we happened to
@@ -1711,7 +1823,7 @@ const stmts = {
  // re-parse re-insert the whole turn history (587k rows across this DB, 75%
  // of the events table, one session holding 68,747 copies of the same turns).
  insertEventAt: db.prepare(
-    "INSERT INTO events (session_id, agent_id, event_type, tool_name, summary, data, created_at) VALUES (?, ?, ?, ?, ?, trim_event_data(?), ?)"
+    "INSERT INTO events (session_id, agent_id, event_type, tool_name, summary, data, created_at) VALUES (?, ?, ?, ?, redact_text(?), trim_event_data(?), ?)"
  ),
  listEvents: db.prepare("SELECT * FROM events ORDER BY created_at DESC, id DESC LIMIT ? OFFSET ?"),
  listEventsBySession: db.prepare(