Files
cairnobs/web/src/lib/api.ts
T
jcoffey-dev 9435115ab7 Phase 3: dashboards and alerting
Saved, shareable multi-panel dashboards (table/line/bar/single-stat
panels via gridstack + uPlot, global + per-panel time range, JSON
export/import) and threshold/absence alert rules with an
ok/pending/firing evaluator and webhook/Slack/PagerDuty delivery.

- New /metadata component: Postgres control-plane store for dashboards,
  panels, notification targets, alert rules/state, and delivery log --
  see docs/phase-3-dashboard-design.md for why ClickHouse's MergeTree
  family isn't a fit for this access pattern (needs real row-level
  locking and read-your-writes consistency).
- api/internal/dashboards: dashboard/panel CRUD, pure -- panel query
  execution stays client-side, reusing the existing /query endpoint.
- New /alerting service: rule/target CRUD, a ticker-driven evaluator
  (claim-then-evaluate concurrency control, transactional-outbox
  delivery, query errors and threshold zero-rows never coerced into a
  false transition) and webhook/Slack/PagerDuty delivery with
  retry/backoff. See docs/phase-3-alerting-design.md for the full
  state-machine design and the four correctness properties it
  implements.
- web: /dashboards and /alerts UIs; cli: sentryctl dashboards/alerts
  list/get/apply, seeding a future Terraform provider's JSON contract.
- hack/alert-load-test: 500 rules against real ClickHouse data, real
  measured results in docs/phase-3-runbook.md.

Five real bugs found by actually running this against a live stack
(documented in the runbook, not just fixed silently): a latent Phase 2
bug where ClickHouse rejected the timestamp format used for
earliest=/latest= queries; a "now" literal token injected into query
text; a GridStack/uPlot layout-timing race; JS's Date.parse being too
lenient to use as a timestamp-detection heuristic; a rule's "enabled"
field silently defaulting to false when omitted; and the evaluator's
claim-batch-size and worker-pool-concurrency defaulting to the same
value, causing 500 concurrently-due rules to take 125s to cycle through
instead of the configured 60s.
2026-08-13 17:29:38 -07:00

243 lines
7.8 KiB
TypeScript

// First real API-client module -- previously each route did its own
// inline fetch(). Introduced in Phase 3 because the surface triples
// (query + dashboards + panels + export/import); still zero-dependency,
// a thin fetch wrapper, not a generated client.
export const apiBase = import.meta.env.VITE_API_BASE_URL ?? 'http://localhost:8080';
export const alertingBase = import.meta.env.VITE_ALERTING_API_BASE_URL ?? 'http://localhost:8081';
export type Language = '' | 'sql' | 'spl';
export type QueryResult = { columns: string[]; rows: unknown[][] };
export type VizType = 'table' | 'line' | 'bar' | 'single_stat' | 'top_n';
export type Panel = {
id: string;
dashboard_id: string;
title: string;
query: string;
query_language: Language;
viz_type: VizType;
viz_config: Record<string, string>;
position_x: number;
position_y: number;
width: number;
height: number;
earliest_override: string | null;
latest_override: string | null;
sort_order: number;
};
export type Dashboard = {
id: string;
name: string;
description: string;
default_earliest: string;
default_latest: string;
created_at: string;
updated_at: string;
panels: Panel[] | null;
};
class ApiError extends Error {}
async function requestFrom<T>(base: string, path: string, init?: RequestInit): Promise<T> {
const res = await fetch(`${base}${path}`, {
headers: { 'Content-Type': 'application/json' },
...init
});
if (!res.ok) {
let message = `request failed with status ${res.status}`;
try {
const body = await res.json();
if (body?.error) message = body.error;
} catch {
// non-JSON error body -- keep the generic message
}
throw new ApiError(message);
}
if (res.status === 204) return undefined as T;
return res.json();
}
function request<T>(path: string, init?: RequestInit): Promise<T> {
return requestFrom(apiBase, path, init);
}
// alerting is a separate service (its own base URL) -- see
// /docs/phase-3-alerting-design.md's component boundary.
function alertingRequest<T>(path: string, init?: RequestInit): Promise<T> {
return requestFrom(alertingBase, path, init);
}
export function runQuery(query: string, language: Language): Promise<QueryResult> {
return request('/query', { method: 'POST', body: JSON.stringify({ query, language }) });
}
export function listDashboards(): Promise<Dashboard[]> {
return request('/dashboards').then((d) => (d as Dashboard[]) ?? []);
}
export function getDashboard(id: string): Promise<Dashboard> {
return request(`/dashboards/${id}`);
}
export function createDashboard(input: {
name: string;
description?: string;
default_earliest?: string;
default_latest?: string;
}): Promise<Dashboard> {
return request('/dashboards', { method: 'POST', body: JSON.stringify(input) });
}
export function updateDashboard(
id: string,
input: { name: string; description?: string; default_earliest?: string; default_latest?: string }
): Promise<Dashboard> {
return request(`/dashboards/${id}`, { method: 'PUT', body: JSON.stringify(input) });
}
export function deleteDashboard(id: string): Promise<void> {
return request(`/dashboards/${id}`, { method: 'DELETE' });
}
export function addPanel(dashboardId: string, panel: Partial<Panel>): Promise<Panel> {
return request(`/dashboards/${dashboardId}/panels`, { method: 'POST', body: JSON.stringify(panel) });
}
export function updatePanel(dashboardId: string, panel: Partial<Panel>): Promise<Panel> {
return request(`/dashboards/${dashboardId}/panels/${panel.id}`, {
method: 'PUT',
body: JSON.stringify(panel)
});
}
export function deletePanel(dashboardId: string, panelId: string): Promise<void> {
return request(`/dashboards/${dashboardId}/panels/${panelId}`, { method: 'DELETE' });
}
export function exportDashboard(id: string): Promise<Dashboard> {
return request(`/dashboards/${id}/export`);
}
export function importDashboard(dashboard: Dashboard): Promise<Dashboard> {
return request('/dashboards/import', { method: 'POST', body: JSON.stringify(dashboard) });
}
// resolveTimeRange applies the override-or-default rule from
// /docs/phase-3-dashboard-design.md's "Time-range mechanics": a panel's
// own earliest/latest override wins if set, otherwise the dashboard's
// default applies.
export function resolveTimeRange(
dashboard: Pick<Dashboard, 'default_earliest' | 'default_latest'>,
panel: Pick<Panel, 'earliest_override' | 'latest_override'>
): { earliest: string; latest: string } {
return {
earliest: panel.earliest_override ?? dashboard.default_earliest,
latest: panel.latest_override ?? dashboard.default_latest
};
}
// injectTimeRange prepends earliest=/latest= as leading base_search
// terms -- works because they're ordinary implicit-AND terms in Phase
// 2's grammar, order-independent. Never used for raw-SQL panels (the
// dashboards API rejects query_language: "sql" on panels entirely, so
// this never has to handle that case).
//
// "now" is a UI-only sentinel (the default_latest value shown in the
// time-range picker), not a token the query language understands --
// time_expr only accepts a quoted absolute timestamp or a "-N unit"
// relative offset (see /docs/query-language-design.md). Emitting a
// literal `latest=now` produces a real compile error ("expected a
// quoted absolute timestamp or a relative offset"), caught by actually
// running this against the live stack. Omitting the latest= clause
// entirely is the query language's own way of saying "no upper bound",
// which is exactly what "now" means here.
export function injectTimeRange(query: string, earliest: string, latest: string): string {
const clauses = [`earliest=${earliest}`];
if (latest && latest !== 'now') clauses.push(`latest=${latest}`);
return `${clauses.join(' ')} ${query}`;
}
// --- alerting ---------------------------------------------------------
export type ConditionType = 'threshold' | 'absence';
export type Comparator = 'gt' | 'gte' | 'lt' | 'lte' | 'eq' | 'ne';
export type NotificationKind = 'webhook' | 'slack' | 'pagerduty';
export type AlertRuleState = 'ok' | 'pending' | 'firing';
export type NotificationTarget = {
id: string;
name: string;
kind: NotificationKind;
webhook_url: string;
};
export type AlertRule = {
id: string;
name: string;
description: string;
query: string;
query_language: Language;
condition_type: ConditionType;
comparator?: Comparator;
threshold_value?: number;
eval_interval_seconds: number;
for_minutes: number;
renotify_interval_minutes?: number;
notification_target_id: string;
enabled: boolean;
state: {
state: AlertRuleState;
last_evaluated_at?: string;
last_eval_status: 'ok' | 'error';
last_error?: string;
last_value?: number;
consecutive_errors: number;
};
};
export type DeliveryLogEntry = {
id: number;
event_type: 'firing' | 'resolved';
status: 'pending' | 'sent' | 'failed' | 'retrying';
attempt_count: number;
last_error?: string;
response_status?: number;
created_at: string;
};
export function listRules(): Promise<AlertRule[]> {
return alertingRequest<AlertRule[]>('/rules').then((r) => r ?? []);
}
export function getRule(id: string): Promise<AlertRule> {
return alertingRequest(`/rules/${id}`);
}
export function createRule(input: Partial<AlertRule>): Promise<AlertRule> {
return alertingRequest('/rules', { method: 'POST', body: JSON.stringify(input) });
}
export function deleteRule(id: string): Promise<void> {
return alertingRequest(`/rules/${id}`, { method: 'DELETE' });
}
export function listDeliveries(ruleId: string): Promise<DeliveryLogEntry[]> {
return alertingRequest<DeliveryLogEntry[]>(`/rules/${ruleId}/deliveries`).then((d) => d ?? []);
}
export function listNotificationTargets(): Promise<NotificationTarget[]> {
return alertingRequest<NotificationTarget[]>('/targets').then((t) => t ?? []);
}
export function createNotificationTarget(input: {
name: string;
kind: NotificationKind;
webhook_url: string;
}): Promise<NotificationTarget> {
return alertingRequest('/targets', { method: 'POST', body: JSON.stringify(input) });
}