fix: add global setErrorHandler and harden getLiveUsercount against ClickHouse failures

server/src/api/analytics/getLiveUsercount.ts

@@ -1,27 +1,51 @@
 import { FastifyReply, FastifyRequest } from "fastify";
+import { z } from "zod";
 import { clickhouse } from "../../db/clickhouse/clickhouse.js";
 import { processResults } from "./utils/utils.js";
 
+const paramsSchema = z.object({
+  siteId: z.coerce.number().int().positive(),
+});
+
+const querySchema = z.object({
+  minutes: z.coerce.number().int().positive().default(5),
+});
+
 export const getLiveUsercount = async (
   req: FastifyRequest<{
     Params: { siteId: string };
     Querystring: { minutes: number };
   }>,
   res: FastifyReply
 ) => {
-  const { siteId } = req.params;
-  const { minutes } = req.query;
+  const paramsValidation = paramsSchema.safeParse(req.params);
+  const queryValidation = querySchema.safeParse(req.query);
+
+  if (!paramsValidation.success || !queryValidation.success) {
+    return res.status(400).send({
+      error: "Invalid request parameters",
+      details: paramsValidation.success ? queryValidation.error : paramsValidation.error,
+    });
+  }
+
+  const { siteId } = paramsValidation.data;
+  const { minutes } = queryValidation.data;
 
-  const query = await clickhouse.query({
-    query: `SELECT COUNT(DISTINCT(session_id)) AS count FROM events WHERE timestamp > now() - interval {minutes:Int32} minute AND site_id = {siteId:Int32}`,
-    format: "JSONEachRow",
-    query_params: {
-      siteId: Number(siteId),
-      minutes: Number(minutes || 5),
-    },
-  });
+  try {
+    const query = await clickhouse.query({
+      query: `SELECT COUNT(DISTINCT(session_id)) AS count FROM events WHERE timestamp > now() - interval {minutes:Int32} minute AND site_id = {siteId:Int32}`,
+      format: "JSONEachRow",
+      query_params: {
+        siteId,
+        minutes,
+      },
+    });
 
-  const result = await processResults<{ count: number }>(query);
+    const result = await processResults<{ count: number }>(query);
 
-  return res.send({ count: result[0].count });
+    return res.send({ count: result[0]?.count ?? 0 });
+  } catch (error) {
+    req.log.error({ err: error, siteId }, "getLiveUsercount: ClickHouse query failed");
+    return res.status(500).send({ error: "Failed to fetch live user count" });
+  }
 };

server/src/index.ts

@@ -188,6 +188,27 @@ const server = Fastify({
   bodyLimit: 10 * 1024 * 1024, // 10MB limit for session replay data
 });
 
+// Global error handler — ensures uncaught route errors are always logged with
+// request context, even on routes that opt out of request logging via
+// `logLevel: "silent"` (e.g. /sites/:siteId/live-user-count, /health). Without
+// this, transient ClickHouse / Postgres outages return 500 to the dashboard
+// with no log signal.
+server.setErrorHandler((error, request, reply) => {
+  request.log.error(
+    { err: error, reqId: request.id, method: request.method, url: request.url },
+    "Unhandled route error"
+  );
+  const statusCode = error.statusCode && error.statusCode >= 400 ? error.statusCode : 500;
+  reply.status(statusCode).send({
+    error: statusCode >= 500 ? "Internal Server Error" : error.message,
+    statusCode,
+  });
+});
+
+server.setNotFoundHandler((request, reply) => {
+  reply.status(404).send({ error: "Not Found", statusCode: 404, path: request.url });
+});
+
 server.register(cors, {
   origin: (_origin, callback) => {
     callback(null, true);