-
-
-
-
-
-
-
-
- Last delivery: {timeAgo(channel.lastDeliveryAt)} ·{" "} - - {channel.lastDeliveryOk ? "OK" : "Failed"} - -
- )} - -{stat.label}
-- {stat.value} -
-Rule Builder
- - {/* Main condition row */} -- {preview} -
-{ch.name}
-- No channels yet. Add one to receive alert notifications. -
- )} -No alert history yet.
- )} - {historyEntries.map((h) => ( -{h.ruleName}
-- {timeAgo(h.firedAt)} · {h.duration} · via {h.channel} -
-No components to report yet.
- )} -{data.detail}
-- This page isn't public yet — enable it from the project switcher to get a shareable link. -
- )} - - {/* Overall status */} -No traffic recorded yet — components appear here once your API receives requests.
- )} -{data.detail}
-Click below to confirm you'd like incident alerts for this status page.
" - f'' - "If you didn't request this, ignore this email.
" - ) - return subject, html - - -def incident_email(title: str, summary: str, status_url: str, unsubscribe_url: str) -> tuple[str, str]: - subject = f"Incident: {title}" - html = ( - f"{title}
" - f"{summary}
" - f'' - f'' - ) - return subject, html diff --git a/apps/api/core/notify.py b/apps/api/core/notify.py deleted file mode 100644 index e60fbdc..0000000 --- a/apps/api/core/notify.py +++ /dev/null @@ -1,62 +0,0 @@ -""" -Alert notification delivery — real HTTP POST to the configured channel. - -Slack / Discord / generic webhooks are just JSON POSTs to an incoming-webhook -URL. PagerDuty uses its Events API v2 (webhook_url holds the routing key). -Returns (ok, detail) and never raises — a delivery failure must not crash the -evaluation loop. -""" -from __future__ import annotations - -import logging - -import httpx - -logger = logging.getLogger(__name__) - -_TIMEOUT = 8.0 - - -async def deliver( - channel_type: str, - webhook_url: str | None, - *, - title: str, - summary: str, - severity: str, -) -> tuple[bool, str]: - if not webhook_url: - return False, "no webhook url configured" - - try: - async with httpx.AsyncClient(timeout=_TIMEOUT) as client: - if channel_type == "slack": - r = await client.post(webhook_url, json={"text": f":rotating_light: *{title}*\n{summary}"}) - elif channel_type == "discord": - r = await client.post(webhook_url, json={"content": f"🚨 **{title}**\n{summary}"}) - elif channel_type == "pagerduty": - r = await client.post( - "https://events.pagerduty.com/v2/enqueue", - json={ - "routing_key": webhook_url, - "event_action": "trigger", - "payload": { - "summary": f"{title} — {summary}"[:1024], - "severity": "critical" if severity == "critical" else "warning", - "source": "liveboard", - }, - }, - ) - else: # generic webhook - r = await client.post( - webhook_url, - json={"title": title, "summary": summary, "severity": severity, "source": "liveboard"}, - ) - - ok = 200 <= r.status_code < 300 - if not ok: - logger.warning("Delivery to %s failed: HTTP %s", channel_type, r.status_code) - return ok, f"HTTP {r.status_code}" - except Exception as exc: # network / timeout / bad url - logger.warning("Delivery to %s errored: %s", channel_type, exc) - return False, str(exc)[:160] diff --git a/apps/api/main.py b/apps/api/main.py index e87f954..ac497b8 100644 --- a/apps/api/main.py +++ b/apps/api/main.py @@ -8,7 +8,7 @@ from fastapi import FastAPI from fastapi.middleware.cors import CORSMiddleware -from api.routes import alerts, ingest, internal, projects, public_status, query, spans +from api.routes import ingest, internal, projects, query from core.config import settings from core.database import close_pool, get_pool from core.redis_client import close_redis @@ -68,9 +68,6 @@ async def lifespan(app: FastAPI): _fastapi.include_router(projects.router) _fastapi.include_router(internal.router) _fastapi.include_router(query.router) -_fastapi.include_router(alerts.router) -_fastapi.include_router(spans.router) -_fastapi.include_router(public_status.router) _fastapi.include_router(sse_router) diff --git a/apps/api/migrations/versions/001_initial.py b/apps/api/migrations/versions/001_initial.py index 47bade0..ca1a75a 100644 --- a/apps/api/migrations/versions/001_initial.py +++ b/apps/api/migrations/versions/001_initial.py @@ -1,4 +1,4 @@ -"""Initial schema — users, projects, events hypertable, spans, incidents +"""Initial schema — users, projects, events hypertable, incidents Revision ID: 001 Revises: @@ -63,24 +63,6 @@ def upgrade() -> None: op.execute("CREATE INDEX IF NOT EXISTS idx_events_project_time ON events(project_id, time DESC);") op.execute("CREATE INDEX IF NOT EXISTS idx_events_route ON events(project_id, route, time DESC);") - # ── Spans (distributed tracing — populated from Phase 6) ───────────────── - op.execute(""" - CREATE TABLE IF NOT EXISTS spans ( - span_id UUID PRIMARY KEY, - trace_id UUID NOT NULL, - parent_id UUID, - project_id UUID NOT NULL, - service_name TEXT, - operation TEXT, - start_time TIMESTAMPTZ, - duration_ms INTEGER, - status_code SMALLINT, - tags JSONB - ); - """) - op.execute("CREATE INDEX IF NOT EXISTS idx_spans_trace ON spans(trace_id);") - op.execute("CREATE INDEX IF NOT EXISTS idx_spans_project ON spans(project_id, start_time DESC);") - # ── 1-minute continuous aggregate ──────────────────────────────────────── op.execute(""" CREATE MATERIALIZED VIEW IF NOT EXISTS events_1min @@ -129,7 +111,6 @@ def upgrade() -> None: def downgrade() -> None: op.execute("DROP TABLE IF EXISTS incidents;") op.execute("DROP MATERIALIZED VIEW IF EXISTS events_1min;") - op.execute("DROP TABLE IF EXISTS spans;") op.execute("DROP TABLE IF EXISTS events;") op.execute("DROP TABLE IF EXISTS projects;") op.execute("DROP TABLE IF EXISTS users;") diff --git a/apps/api/migrations/versions/002_alert_rules.py b/apps/api/migrations/versions/002_alert_rules.py deleted file mode 100644 index 82d9ba2..0000000 --- a/apps/api/migrations/versions/002_alert_rules.py +++ /dev/null @@ -1,58 +0,0 @@ -"""Add alert_rules and alert_history tables - -Revision ID: 002 -Revises: 001 -Create Date: 2026-06-13 -""" - -from alembic import op - -revision = "002" -down_revision = "001" -branch_labels = None -depends_on = None - - -def upgrade() -> None: - op.execute(""" - CREATE TABLE IF NOT EXISTS alert_rules ( - id UUID PRIMARY KEY DEFAULT gen_random_uuid(), - project_id UUID NOT NULL REFERENCES projects(id) ON DELETE CASCADE, - name TEXT NOT NULL, - metric TEXT NOT NULL, - operator TEXT NOT NULL CHECK (operator IN ('>', '<', '=', '!=')), - threshold FLOAT NOT NULL, - window_min INTEGER NOT NULL DEFAULT 5, - severity TEXT NOT NULL DEFAULT 'warning' - CHECK (severity IN ('critical', 'warning', 'info')), - channel TEXT NOT NULL DEFAULT '', - status TEXT NOT NULL DEFAULT 'ok' - CHECK (status IN ('firing', 'ok', 'pending')), - enabled BOOLEAN NOT NULL DEFAULT TRUE, - last_triggered TIMESTAMPTZ, - created_at TIMESTAMPTZ DEFAULT now(), - updated_at TIMESTAMPTZ DEFAULT now() - ); - """) - op.execute("CREATE INDEX IF NOT EXISTS idx_alert_rules_project ON alert_rules(project_id);") - - op.execute(""" - CREATE TABLE IF NOT EXISTS alert_history ( - id UUID PRIMARY KEY DEFAULT gen_random_uuid(), - project_id UUID NOT NULL, - rule_id UUID REFERENCES alert_rules(id) ON DELETE SET NULL, - rule_name TEXT NOT NULL, - fired_at TIMESTAMPTZ NOT NULL DEFAULT now(), - resolved_at TIMESTAMPTZ, - channel TEXT NOT NULL DEFAULT '' - ); - """) - op.execute( - "CREATE INDEX IF NOT EXISTS idx_alert_history_project " - "ON alert_history(project_id, fired_at DESC);" - ) - - -def downgrade() -> None: - op.execute("DROP TABLE IF EXISTS alert_history;") - op.execute("DROP TABLE IF EXISTS alert_rules;") diff --git a/apps/api/migrations/versions/003_multitenancy.py b/apps/api/migrations/versions/002_multitenancy.py similarity index 96% rename from apps/api/migrations/versions/003_multitenancy.py rename to apps/api/migrations/versions/002_multitenancy.py index 601a2f2..e1e5ed1 100644 --- a/apps/api/migrations/versions/003_multitenancy.py +++ b/apps/api/migrations/versions/002_multitenancy.py @@ -1,7 +1,7 @@ """Multi-tenancy: organizations, memberships, projects.org_id -Revision ID: 003 -Revises: 002 +Revision ID: 002 +Revises: 001 Create Date: 2026-07-18 Introduces the tenancy identity layer: @@ -10,15 +10,15 @@ projects.org_id — every project now belongs to exactly one org Telemetry tables already carry project_id, so an org owns its data -transitively (org -> projects -> events/spans/incidents). This migration +transitively (org -> projects -> events/incidents). This migration only adds the *ownership* layer; read-time authorization is enforced in the app (Phase 8.2) and, later, Postgres RLS (Phase 8.4). """ from alembic import op -revision = "003" -down_revision = "002" +revision = "002" +down_revision = "001" branch_labels = None depends_on = None diff --git a/apps/api/migrations/versions/004_api_keys.py b/apps/api/migrations/versions/003_api_keys.py similarity index 96% rename from apps/api/migrations/versions/004_api_keys.py rename to apps/api/migrations/versions/003_api_keys.py index 6defb75..c211cf8 100644 --- a/apps/api/migrations/versions/004_api_keys.py +++ b/apps/api/migrations/versions/003_api_keys.py @@ -1,7 +1,7 @@ """API keys table — multiple keys per project, rotation + revocation -Revision ID: 004 -Revises: 003 +Revision ID: 003 +Revises: 002 Create Date: 2026-07-18 Moves API-key auth from the single `projects.api_key` column to a dedicated @@ -11,8 +11,8 @@ from alembic import op -revision = "004" -down_revision = "003" +revision = "003" +down_revision = "002" branch_labels = None depends_on = None diff --git a/apps/api/migrations/versions/005_rls.py b/apps/api/migrations/versions/004_rls.py similarity index 88% rename from apps/api/migrations/versions/005_rls.py rename to apps/api/migrations/versions/004_rls.py index b153919..bad79a8 100644 --- a/apps/api/migrations/versions/005_rls.py +++ b/apps/api/migrations/versions/004_rls.py @@ -1,7 +1,7 @@ """Row-Level Security backstop for tenant isolation -Revision ID: 005 -Revises: 004 +Revision ID: 004 +Revises: 003 Create Date: 2026-07-18 Defense-in-depth: even if an app query forgets its `WHERE project_id = …`, @@ -12,7 +12,7 @@ • Dashboard READ connections `SET ROLE dashboard_reader` + set the `app.project_id` GUC (see api.deps.scoped_conn). Non-owner ⇒ RLS is enforced ⇒ only rows for the active project are visible. - • The owner (ingest worker, alert writes) is NOT forced under RLS, so + • The owner (ingest worker, incident writes) is NOT forced under RLS, so cross-project writes/aggregation keep working. A read that forgets to set `app.project_id` sees ZERO rows (fail closed). @@ -20,12 +20,12 @@ from alembic import op -revision = "005" -down_revision = "004" +revision = "004" +down_revision = "003" branch_labels = None depends_on = None -_TABLES = ["events", "spans", "incidents", "alert_rules", "alert_history"] +_TABLES = ["events", "incidents"] def upgrade() -> None: @@ -46,7 +46,7 @@ def upgrade() -> None: op.execute(f"GRANT SELECT ON {t} TO dashboard_reader;") op.execute(f"ALTER TABLE {t} ENABLE ROW LEVEL SECURITY;") # Scoped SELECT policy — applies only to dashboard_reader; the owner - # (worker/ingest/alert writes) is unaffected. + # (worker/ingest/incident writes) is unaffected. op.execute(f""" CREATE POLICY {t}_tenant_isolation ON {t} FOR SELECT diff --git a/apps/api/migrations/versions/006_alert_channels.py b/apps/api/migrations/versions/006_alert_channels.py deleted file mode 100644 index 9534c41..0000000 --- a/apps/api/migrations/versions/006_alert_channels.py +++ /dev/null @@ -1,50 +0,0 @@ -"""Alert delivery channels (real Slack/Discord/PagerDuty/webhook targets) - -Revision ID: 006 -Revises: 005 -Create Date: 2026-07-19 - -Backs the alerts "Channels" tab with real, per-project destinations. Alert -rules reference a channel by name (alert_rules.channel); the evaluation worker -looks the channel up here to get its webhook URL and delivers the notification. -""" - -from alembic import op - -revision = "006" -down_revision = "005" -branch_labels = None -depends_on = None - - -def upgrade() -> None: - op.execute(""" - CREATE TABLE IF NOT EXISTS alert_channels ( - id UUID PRIMARY KEY DEFAULT gen_random_uuid(), - project_id UUID NOT NULL REFERENCES projects(id) ON DELETE CASCADE, - type TEXT NOT NULL CHECK (type IN ('slack', 'discord', 'pagerduty', 'webhook')), - name TEXT NOT NULL, - webhook_url TEXT, - enabled BOOLEAN NOT NULL DEFAULT TRUE, - last_delivery_at TIMESTAMPTZ, - last_delivery_ok BOOLEAN, - created_at TIMESTAMPTZ DEFAULT now() - ); - """) - op.execute("CREATE INDEX IF NOT EXISTS idx_alert_channels_project ON alert_channels(project_id);") - - # RLS backstop — dashboard reads (dashboard_reader role) are scoped to the - # active project via app.project_id, same as the other tenant tables. - op.execute("GRANT SELECT ON alert_channels TO dashboard_reader;") - op.execute("ALTER TABLE alert_channels ENABLE ROW LEVEL SECURITY;") - op.execute(""" - CREATE POLICY alert_channels_tenant_isolation ON alert_channels - FOR SELECT - TO dashboard_reader - USING (project_id = NULLIF(current_setting('app.project_id', true), '')::uuid); - """) - - -def downgrade() -> None: - op.execute("DROP POLICY IF EXISTS alert_channels_tenant_isolation ON alert_channels;") - op.execute("DROP TABLE IF EXISTS alert_channels;") diff --git a/apps/api/migrations/versions/007_public_status_page.py b/apps/api/migrations/versions/007_public_status_page.py deleted file mode 100644 index 3204d55..0000000 --- a/apps/api/migrations/versions/007_public_status_page.py +++ /dev/null @@ -1,36 +0,0 @@ -"""Public per-project status page (opt-in) - -Revision ID: 007 -Revises: 006 -Create Date: 2026-07-20 - -Lets a project owner publish a logged-out-visitable status page at -/status/{public_slug}, scoped to exactly that project. Off by default — -`status_page_enabled` must be flipped on (see /v1/internal/projects/{id}/status-page) -before the public route will serve anything for a project. -""" - -from alembic import op - -revision = "007" -down_revision = "006" -branch_labels = None -depends_on = None - - -def upgrade() -> None: - op.execute("ALTER TABLE projects ADD COLUMN IF NOT EXISTS public_slug TEXT;") - op.execute( - "ALTER TABLE projects ADD COLUMN IF NOT EXISTS status_page_enabled " - "BOOLEAN NOT NULL DEFAULT FALSE;" - ) - op.execute( - "CREATE UNIQUE INDEX IF NOT EXISTS idx_projects_public_slug " - "ON projects(public_slug) WHERE public_slug IS NOT NULL;" - ) - - -def downgrade() -> None: - op.execute("DROP INDEX IF EXISTS idx_projects_public_slug;") - op.execute("ALTER TABLE projects DROP COLUMN IF EXISTS status_page_enabled;") - op.execute("ALTER TABLE projects DROP COLUMN IF EXISTS public_slug;") diff --git a/apps/api/migrations/versions/008_status_subscribers.py b/apps/api/migrations/versions/008_status_subscribers.py deleted file mode 100644 index 1f2a092..0000000 --- a/apps/api/migrations/versions/008_status_subscribers.py +++ /dev/null @@ -1,42 +0,0 @@ -"""Status page incident-alert subscribers (double opt-in by email) - -Revision ID: 008 -Revises: 007 -Create Date: 2026-07-20 - -Backs the public status page's "Get Incident Alerts" form. A row is created -on subscribe with `confirmed_at = NULL`; the anomaly worker only emails rows -where `confirmed_at IS NOT NULL AND unsubscribed_at IS NULL`, so an address is -never emailed until it's clicked its confirmation link. -""" - -from alembic import op - -revision = "008" -down_revision = "007" -branch_labels = None -depends_on = None - - -def upgrade() -> None: - op.execute(""" - CREATE TABLE IF NOT EXISTS status_subscribers ( - id UUID PRIMARY KEY DEFAULT gen_random_uuid(), - project_id UUID NOT NULL REFERENCES projects(id) ON DELETE CASCADE, - email TEXT NOT NULL, - confirm_token TEXT UNIQUE NOT NULL, - confirmed_at TIMESTAMPTZ, - unsubscribed_at TIMESTAMPTZ, - created_at TIMESTAMPTZ DEFAULT now(), - UNIQUE (project_id, email) - ); - """) - op.execute( - "CREATE INDEX IF NOT EXISTS idx_status_subscribers_active " - "ON status_subscribers(project_id) " - "WHERE confirmed_at IS NOT NULL AND unsubscribed_at IS NULL;" - ) - - -def downgrade() -> None: - op.execute("DROP TABLE IF EXISTS status_subscribers;") diff --git a/apps/api/requirements.txt b/apps/api/requirements.txt index eea978a..360640a 100644 --- a/apps/api/requirements.txt +++ b/apps/api/requirements.txt @@ -18,9 +18,5 @@ python-socketio[asyncio_client]==5.11.4 # AI (Phase 5) cerebras-cloud-sdk==1.19.0 -# Auth -python-jose[cryptography]==3.3.0 -passlib[bcrypt]==1.7.4 - -# HTTP client (smoke test + SDK calls) +# HTTP client (smoke test) httpx==0.27.2 diff --git a/apps/api/worker/alerts.py b/apps/api/worker/alerts.py deleted file mode 100644 index a685361..0000000 --- a/apps/api/worker/alerts.py +++ /dev/null @@ -1,150 +0,0 @@ -""" -Alert-rule evaluation worker. - -Every 30 s, for each ENABLED rule: - 1. Compute its metric over the rule's window from the events table. - 2. Compare against the threshold using the rule's operator. - 3. State machine: - - breach + not already firing -> status='firing', last_triggered=now, - insert alert_history row, deliver to the rule's channel - - no breach + currently firing -> status='ok', resolve the open history row - Owner DB connection (bypasses RLS); the worker sees every project's rules. -""" -from __future__ import annotations - -import asyncio -import logging - -from core.database import get_pool -from core.notify import deliver - -logger = logging.getLogger(__name__) - -_INTERVAL = 30 # seconds - -_METRIC = { - "error_rate": ("Error rate", "%"), - "p99_latency": ("p99 latency", "ms"), - "requests_per_min": ("Requests/min", ""), -} -_OP_WORD = {">": "exceeds", "<": "drops below", "=": "equals", "!=": "differs from"} - - -def _compare(value: float, operator: str, threshold: float) -> bool: - if operator == ">": - return value > threshold - if operator == "<": - return value < threshold - if operator == "=": - return abs(value - threshold) < 1e-6 - if operator == "!=": - return abs(value - threshold) >= 1e-6 - return False - - -async def _metric_value(conn, project_id, metric: str, window_min: int) -> float | None: - """Return the current metric value over the window, or None if not evaluable.""" - row = await conn.fetchrow( - """ - SELECT - COUNT(*)::int AS total, - COALESCE(100.0 * SUM((status_code >= 400)::int) / NULLIF(COUNT(*), 0), 0)::float AS error_rate, - COALESCE(percentile_cont(0.99) WITHIN GROUP (ORDER BY duration_ms), 0)::float AS p99 - FROM events - WHERE project_id = $1 AND time > now() - ($2 * interval '1 minute') - """, - project_id, - window_min, - ) - total = row["total"] or 0 - if metric == "requests_per_min": - return total / max(window_min, 1) - if total == 0: - return None # can't judge error rate / latency with no traffic - if metric == "error_rate": - return float(row["error_rate"]) - if metric == "p99_latency": - return float(row["p99"]) - return None # unsupported metric (e.g. apdex) — skip - - -async def _deliver_for_rule(conn, project_id, rule, title: str, summary: str) -> None: - ch = await conn.fetchrow( - """ - SELECT id, type, webhook_url FROM alert_channels - WHERE project_id = $1 AND name = $2 AND enabled = TRUE - LIMIT 1 - """, - project_id, - rule["channel"], - ) - if ch is None: - return - ok, _detail = await deliver(ch["type"], ch["webhook_url"], title=title, summary=summary, severity=rule["severity"]) - await conn.execute( - "UPDATE alert_channels SET last_delivery_at = now(), last_delivery_ok = $2 WHERE id = $1", - ch["id"], - ok, - ) - - -async def _evaluate_all() -> None: - pool = await get_pool() - async with pool.acquire() as conn: - rules = await conn.fetch( - """ - SELECT id, project_id, name, metric, operator, threshold, window_min, - severity, channel, status - FROM alert_rules - WHERE enabled = TRUE - """ - ) - for rule in rules: - value = await _metric_value(conn, rule["project_id"], rule["metric"], rule["window_min"]) - if value is None: - continue - - breach = _compare(value, rule["operator"], rule["threshold"]) - label, unit = _METRIC.get(rule["metric"], (rule["metric"], "")) - op_word = _OP_WORD.get(rule["operator"], rule["operator"]) - - if breach and rule["status"] != "firing": - title = f"{label} {op_word} {rule['threshold']:g}{unit}" - summary = ( - f"{label} hit {value:.1f}{unit} over the last {rule['window_min']}m " - f"(threshold {rule['operator']} {rule['threshold']:g}{unit})." - ) - await conn.execute( - "UPDATE alert_rules SET status='firing', last_triggered=now(), updated_at=now() WHERE id=$1", - rule["id"], - ) - await conn.execute( - """ - INSERT INTO alert_history (project_id, rule_id, rule_name, fired_at, channel) - VALUES ($1, $2, $3, now(), $4) - """, - rule["project_id"], rule["id"], rule["name"], rule["channel"], - ) - await _deliver_for_rule(conn, rule["project_id"], rule, title, summary) - logger.info("Alert FIRING rule=%s project=%s value=%.1f", rule["name"], rule["project_id"], value) - - elif not breach and rule["status"] == "firing": - await conn.execute( - "UPDATE alert_rules SET status='ok', updated_at=now() WHERE id=$1", - rule["id"], - ) - await conn.execute( - "UPDATE alert_history SET resolved_at=now() WHERE rule_id=$1 AND resolved_at IS NULL", - rule["id"], - ) - logger.info("Alert RESOLVED rule=%s project=%s", rule["name"], rule["project_id"]) - - -async def run() -> None: - logger.info("Alert evaluator started — checking every %d s", _INTERVAL) - while True: - try: - await _evaluate_all() - except Exception: - logger.exception("Alert evaluation cycle failed") - await asyncio.sleep(_INTERVAL) diff --git a/apps/api/worker/anomaly.py b/apps/api/worker/anomaly.py index dda0b85..e8ff03c 100644 --- a/apps/api/worker/anomaly.py +++ b/apps/api/worker/anomaly.py @@ -21,7 +21,6 @@ from core.config import settings from core.database import get_pool -from core.email import incident_email, send as send_email from core.redis_client import get_redis logger = logging.getLogger(__name__) @@ -241,37 +240,6 @@ async def _handle_anomaly( await redis.publish(f"incidents:{project_id}", json.dumps(payload)) logger.info("Incident created id=%s project=%s", row["id"], project_id) - await _notify_subscribers(conn, project_id, title, summary) - - -async def _notify_subscribers(conn, project_id: str, title: str, summary: str) -> None: - """ - Email confirmed, active status-page subscribers about a new incident. - Best-effort — a delivery failure must not break the anomaly loop, same as - the alert-channel worker's notify.deliver. - """ - rows = await conn.fetch( - """ - SELECT s.email, s.confirm_token, p.public_slug - FROM status_subscribers s - JOIN projects p ON p.id = s.project_id - WHERE s.project_id = $1 AND s.confirmed_at IS NOT NULL AND s.unsubscribed_at IS NULL - """, - project_id, - ) - if not rows or not rows[0]["public_slug"]: - return - - status_url = f"{settings.public_app_url}/status/{rows[0]['public_slug']}" - for row in rows: - unsubscribe_url = f"{settings.public_app_url}/status/unsubscribe/{row['confirm_token']}" - subject, html = incident_email(title, summary, status_url, unsubscribe_url) - ok, detail = await send_email(row["email"], subject, html) - if not ok: - logger.warning("Incident email to %s failed: %s", row["email"], detail) - - -# ─── Cerebras summariser ───────────────────────────────────────────────────── async def _generate_summary( metric: str, diff --git a/apps/api/worker/main.py b/apps/api/worker/main.py index 19ed504..164571b 100644 --- a/apps/api/worker/main.py +++ b/apps/api/worker/main.py @@ -4,9 +4,10 @@ Run as: python -m worker.main Docker: CMD ["python", "-m", "worker.main"] -Two concurrent tasks: +Three concurrent tasks: aggregator — XREADGROUP → TimescaleDB COPY (at-least-once delivery) metrics — query last-minute stats per project → Redis pub/sub every 1 s + anomaly — rolling z-score check every 60 s → AI incident summary """ import asyncio @@ -14,7 +15,6 @@ import signal from worker.aggregator import run as run_aggregator -from worker.alerts import run as run_alerts from worker.anomaly import run as run_anomaly from worker.metrics import run as run_metrics @@ -30,20 +30,18 @@ async def main() -> None: agg_task = asyncio.create_task(run_aggregator(), name="aggregator") metrics_task = asyncio.create_task(run_metrics(), name="metrics-publisher") anomaly_task = asyncio.create_task(run_anomaly(), name="anomaly-detector") - alerts_task = asyncio.create_task(run_alerts(), name="alert-evaluator") def _shutdown(sig: signal.Signals) -> None: logger.info("Received %s, stopping worker", sig.name) agg_task.cancel() metrics_task.cancel() anomaly_task.cancel() - alerts_task.cancel() for sig in (signal.SIGINT, signal.SIGTERM): loop.add_signal_handler(sig, _shutdown, sig) try: - await asyncio.gather(agg_task, metrics_task, anomaly_task, alerts_task, return_exceptions=True) + await asyncio.gather(agg_task, metrics_task, anomaly_task, return_exceptions=True) except asyncio.CancelledError: pass diff --git a/components/alerts/rule-list.tsx b/components/alerts/rule-list.tsx deleted file mode 100644 index 2d52696..0000000 --- a/components/alerts/rule-list.tsx +++ /dev/null @@ -1,137 +0,0 @@ -"use client"; - -import { useEffect, useState } from "react"; -import { Zap, Bell, Info } from "lucide-react"; -import { Badge } from "@/components/ui/badge"; -import { timeAgo, cn } from "@/lib/utils"; -import type { AlertRule } from "@/lib/types"; - -interface RuleListProps { - rules: AlertRule[]; - onToggle?: (id: string, enabled: boolean) => void; -} - -const METRIC_LABELS: Record{payload[0].payload.label}
-{payload[0].value}ms
-{m.label}
-p99 Latency — 24h
-{endpoint.route}
-{stat.label}
-- {stat.value} -
-Latency Distribution
-{p.label}
-- {p.value} -
-Top Errors
- {endpoint.errorRate > 0 ? ( -No errors in selected window
- )} -
- );
-}
diff --git a/components/landing/feature-grid.tsx b/components/landing/feature-grid.tsx
index 65cc6aa..cc7a854 100644
--- a/components/landing/feature-grid.tsx
+++ b/components/landing/feature-grid.tsx
@@ -1,6 +1,5 @@
import { RevealOnScroll } from "@/components/landing/reveal-on-scroll";
import { ScreenshotFrame } from "@/components/landing/screenshot-frame";
-import { ALERT_CHANNELS, ChannelMarkIcon } from "@/components/landing/channel-marks";
import { InstallSnippet } from "@/components/landing/install-snippet";
export function FeatureGrid() {
@@ -11,7 +10,7 @@ export function FeatureGrid() {
Every request, instrumented.
- Distributed traces, per-endpoint health scores, and alerts routed to the tools your team already uses. + Live request metrics, per-endpoint health scores, and AI incident summaries — from one line of middleware.
@@ -22,14 +21,14 @@ export function FeatureGrid() {- Flame graphs and service maps for every request, down to the individual span. Trace IDs propagate across all five SDK adapters automatically. + Request volume, response codes, and a tailing log of every call — pushed to the browser over Socket.io and SSE as the traffic happens, not on a refresh timer.
- Threshold rules on error rate or latency, routed to the channel your on-call rotation actually watches. + A rolling z-score over the last 24 hours of error rate and p99 latency decides what is actually unusual for your API — no thresholds to tune, no alert fatigue.
- Open source API observability. Traces, alerts, and AI incident summaries from one line of middleware. + Open source API observability. Live metrics, error tracking, and AI incident summaries from one line of middleware.
A rolling z-score model flags anomalies in error rate and p99 latency, then writes a plain-English summary automatically. diff --git a/components/landing/navbar.tsx b/components/landing/navbar.tsx index 4df9363..2086ea9 100644 --- a/components/landing/navbar.tsx +++ b/components/landing/navbar.tsx @@ -10,7 +10,6 @@ import { primaryButtonClass } from "@/components/landing/button-styles"; const NAV_LINKS = [ { label: "Features", href: "#features" }, { label: "Detection", href: "#detection" }, - { label: "Status page", href: "#status" }, ]; const GITHUB_URL = "https://github.com/ryzrr/liveboard"; diff --git a/components/landing/stats-section.tsx b/components/landing/stats-section.tsx index fc10266..a7e9d06 100644 --- a/components/landing/stats-section.tsx +++ b/components/landing/stats-section.tsx @@ -3,8 +3,7 @@ import { AnimatedCounter } from "@/components/landing/animated-counter"; const STATS = [ { target: 2, label: "official SDKs" }, - { target: 5, label: "framework adapters" }, - { target: 4, label: "alert channels" }, + { target: 3, label: "worker loops" }, ]; export function StatsSection() { @@ -15,11 +14,11 @@ export function StatsSection() { Built to be adopted in an afternoon.
- No agents, no sidecars, no new infrastructure to run. Two SDKs, five adapters, and a status page for free. + No agents, no sidecars, no new infrastructure to run. Two SDKs, one line of middleware, and Docker Compose.
- Generated from the same data you already collect: uptime, incident timelines, and component health, published automatically. -
-Public status page
-- Anyone with the link can view it — no account needed. -
-{statusError}
} -Danger zone
@@ -208,7 +126,7 @@ export function ProjectSettingsCard({ project }: Props) {This permanently deletes {project.name} and - all its events, traces, alert rules, and API keys. Type the project name to confirm. + all its events, incidents, and API keys. Type the project name to confirm.
{ - const id = setInterval(() => router.refresh(), intervalMs); - return () => clearInterval(id); - }, [router, intervalMs]); - return null; -} diff --git a/components/status/incident-card.tsx b/components/status/incident-card.tsx deleted file mode 100644 index 2622928..0000000 --- a/components/status/incident-card.tsx +++ /dev/null @@ -1,123 +0,0 @@ -"use client"; - -import { useState, useMemo } from "react"; -import { ChevronDown } from "lucide-react"; -import { Badge } from "@/components/ui/badge"; -import { cn, timeAgo } from "@/lib/utils"; -import type { Incident } from "@/lib/types"; - -type TimelineStep = "investigating" | "identified" | "monitoring" | "resolved"; - -const STEP_COLORS: RecordTimeline
-{note}
-- {lastChecked ? `Last checked: ${lastChecked} · ` : ""}Updates in real time -
-- Subscribe to receive email notifications for incidents and status updates. -
- - {state === "sent" ? ( -Check your email to confirm the subscription.
- ) : ( -Something went wrong — try again.
- )} -{service.name}
-{service.uptime90d.toFixed(2)}%
-uptime (90d)
-Service Map
-{field.label}
-- {field.value} -
-Tags
-