max / makenotwork

3.7 KB · 117 lines History Blame Raw

1	//! Inbound infra-alert log (see migration `169_admin_alerts.sql`).
2	//!
3	//! Backs the `POST /api/internal/alerts` ingestion endpoint: external
4	//! monitoring agents (PoM, MT) push ops alerts here, each persisted as one row
5	//! and emailed to the operator. Admin-routed only -- never creator-facing.
6
7	use serde::Deserialize;
8	use sqlx::PgPool;
9
10	/// Alert domain an agent may report. Closed set, deserialized straight from the
11	/// wire (`#[serde(rename_all = "snake_case")]`), so an unknown value is rejected
12	/// when the request body is parsed, there is no runtime membership check.
13	/// Extending the taxonomy is a variant here plus its `as_str`, not a migration
14	/// (the column is TEXT).
15	///
16	/// These are domain-level, not per-event: PoM's finer categories (tls expiry vs
17	/// tls error vs tls recovery) all fold onto `Tls`, with the failure/recovery
18	/// distinction carried by [`AlertSeverity`] instead. Kept small so the operator
19	/// log filters by system, not by every sub-condition.
20	#[derive(Debug, Clone, Copy, Deserialize)]
21	#[serde(rename_all = "snake_case")]
22	pub enum AlertKind {
23	Health,
24	Tls,
25	Dns,
26	Whois,
27	Latency,
28	Cors,
29	Backup,
30	Peer,
31	Route,
32	Scan,
33	Monitoring,
34	}
35
36	impl AlertKind {
37	/// Canonical wire/storage token. Matches the `serde(rename_all)` mapping.
38	pub fn as_str(self) -> &'static str {
39	match self {
40	Self::Health => "health",
41	Self::Tls => "tls",
42	Self::Dns => "dns",
43	Self::Whois => "whois",
44	Self::Latency => "latency",
45	Self::Cors => "cors",
46	Self::Backup => "backup",
47	Self::Peer => "peer",
48	Self::Route => "route",
49	Self::Scan => "scan",
50	Self::Monitoring => "monitoring",
51	}
52	}
53	}
54
55	/// Alert urgency, most to least. Closed set, deserialized like [`AlertKind`].
56	#[derive(Debug, Clone, Copy, Deserialize)]
57	#[serde(rename_all = "snake_case")]
58	pub enum AlertSeverity {
59	Critical,
60	Warning,
61	Info,
62	}
63
64	impl AlertSeverity {
65	/// Canonical wire/storage token.
66	pub fn as_str(self) -> &'static str {
67	match self {
68	Self::Critical => "critical",
69	Self::Warning => "warning",
70	Self::Info => "info",
71	}
72	}
73	}
74
75	/// A validated inbound alert, ready to persist.
76	pub struct NewAlert<'a> {
77	pub source: &'a str,
78	pub kind: AlertKind,
79	pub severity: AlertSeverity,
80	pub title: &'a str,
81	pub body: &'a str,
82	pub dedup_key: Option<&'a str>,
83	pub details: Option<&'a serde_json::Value>,
84	}
85
86	/// Insert an ingested alert and return its generated id. `emailed` starts false;
87	/// the handler flips it via [`mark_emailed`] after the notification is sent so a
88	/// mail failure stays visible in the log.
89	#[tracing::instrument(skip_all, fields(source = alert.source, kind = alert.kind.as_str(), severity = alert.severity.as_str()))]
90	pub async fn insert_alert(pool: &PgPool, alert: &NewAlert<'_>) -> Result<i64, sqlx::Error> {
91	let id: i64 = sqlx::query_scalar(
92	"INSERT INTO admin_alerts (source, kind, severity, title, body, dedup_key, details)
93	VALUES ($1, $2, $3, $4, $5, $6, $7)
94	RETURNING id",
95	)
96	.bind(alert.source)
97	.bind(alert.kind.as_str())
98	.bind(alert.severity.as_str())
99	.bind(alert.title)
100	.bind(alert.body)
101	.bind(alert.dedup_key)
102	.bind(alert.details)
103	.fetch_one(pool)
104	.await?;
105	Ok(id)
106	}
107
108	/// Mark an alert row as successfully emailed to the operator.
109	#[tracing::instrument(skip_all)]
110	pub async fn mark_emailed(pool: &PgPool, id: i64) -> Result<(), sqlx::Error> {
111	sqlx::query("UPDATE admin_alerts SET emailed = true WHERE id = $1")
112	.bind(id)
113	.execute(pool)
114	.await?;
115	Ok(())
116	}
117