@thotischner/observability-mcp 1.4.1 → 1.6.0

package/dist/tools/handlers.test.js

@@ -3,6 +3,8 @@ import assert from "node:assert/strict";
 import { ConnectorRegistry } from "../connectors/registry.js";
 import { listSourcesHandler } from "./list-sources.js";
 import { listServicesHandler } from "./list-services.js";
+import { detectAnomaliesHandler } from "./detect-anomalies.js";
+import { getServiceHealthHandler } from "./get-service-health.js";
 // --- Mock Connector ---
 function createMockConnector(overrides) {
     return {
@@ -136,3 +138,105 @@ describe("listServicesHandler", () => {
         assert.equal(data.total, 0);
     });
 });
+describe("detectAnomaliesHandler — A5 memory/OOM coverage", () => {
+    const flatMemory = () => ({
+        source: "prom1", service: "payment-service", metric: "memory", unit: "bytes",
+        values: Array.from({ length: 40 }, (_, i) => ({
+            timestamp: new Date(Date.now() - (40 - i) * 9000).toISOString(),
+            value: 1.3e8 + (i % 3) * 1e6, // noisy, no trend → no metric anomaly
+        })),
+        summary: { current: 1.3e8, average: 1.3e8, min: 1.28e8, max: 1.33e8, trend: "stable" },
+    });
+    it("scans the memory metric (now in KEY_METRICS)", async () => {
+        const requested = [];
+        const reg = createRegistryWithMocks([
+            createMockConnector({
+                name: "prom1", type: "prometheus", signalType: "metrics",
+                listServices: async () => [{ name: "payment-service", source: "prom1", signalType: "metrics" }],
+                queryMetrics: async ({ metric }) => {
+                    requested.push(metric);
+                    return flatMemory();
+                },
+            }),
+        ]);
+        await detectAnomaliesHandler(reg, {});
+        assert.ok(requested.includes("memory"), `memory not scanned; got ${requested.join(",")}`);
+    });
+    it("flags a warn-level OOM log pattern below the error-rate gate", async () => {
+        const reg = createRegistryWithMocks([
+            createMockConnector({
+                name: "prom1", type: "prometheus", signalType: "metrics",
+                listServices: async () => [{ name: "payment-service", source: "prom1", signalType: "metrics" }],
+                queryMetrics: async () => flatMemory(),
+            }),
+            createMockConnector({
+                name: "loki1", type: "loki", signalType: "logs",
+                queryLogs: async () => ({
+                    source: "loki1", service: "payment-service", entries: [],
+                    // Only 4 warn-level lines: errorCount below the >5 gate, ratio tiny.
+                    summary: {
+                        total: 800, errorCount: 4, warnCount: 4,
+                        topPatterns: ["OutOfMemoryWarning: heap usage exceeding threshold (4x)"],
+                    },
+                }),
+            }),
+        ]);
+        const result = await detectAnomaliesHandler(reg, {});
+        const data = JSON.parse(result.content[0].text);
+        const crit = data.anomalies.find((a) => a.metric === "log_critical_pattern");
+        assert.ok(crit, "expected a log_critical_pattern anomaly for the OOM warning");
+        assert.equal(crit.service, "payment-service");
+        assert.equal(crit.severity, "high");
+        assert.equal(data.rootCause.ranked[0].service, "payment-service");
+        assert.notEqual(data.summary, "All services healthy — no anomalies detected.");
+    });
+    it("does not flag benign warn patterns", async () => {
+        const reg = createRegistryWithMocks([
+            createMockConnector({
+                name: "prom1", type: "prometheus", signalType: "metrics",
+                listServices: async () => [{ name: "order-service", source: "prom1", signalType: "metrics" }],
+                queryMetrics: async () => flatMemory(),
+            }),
+            createMockConnector({
+                name: "loki1", type: "loki", signalType: "logs",
+                queryLogs: async () => ({
+                    source: "loki1", service: "order-service", entries: [],
+                    summary: { total: 500, errorCount: 1, warnCount: 2, topPatterns: ["cache miss for key user:42"] },
+                }),
+            }),
+        ]);
+        const result = await detectAnomaliesHandler(reg, {});
+        const data = JSON.parse(result.content[0].text);
+        assert.equal(data.anomalies.length, 0);
+    });
+});
+describe("getServiceHealthHandler — one-sided latency (regression)", () => {
+    const series = (vals) => ({
+        source: "prom1", service: "payment-service", metric: "x", unit: "",
+        values: vals.map((v, i) => ({ timestamp: new Date(Date.now() - (vals.length - i) * 9000).toISOString(), value: v })),
+        summary: { current: vals[vals.length - 1], average: vals[0], min: Math.min(...vals), max: Math.max(...vals), trend: "falling" },
+    });
+    it("a DECREASING latency_p99 is NOT flagged as an anomaly", async () => {
+        const reg = new ConnectorRegistry();
+        const mock = {
+            connect: async () => { }, disconnect: async () => { },
+            healthCheck: async () => ({ status: "up", latencyMs: 1 }),
+            getDefaultMetrics: () => [], getMetrics: () => [],
+            listServices: async () => [{ name: "payment-service", source: "prom1", signalType: "metrics" }],
+            name: "prom1", type: "prometheus", signalType: "metrics",
+            queryMetrics: async ({ metric }) => {
+                if (metric === "latency_p99")
+                    return series(Array.from({ length: 30 }, (_, i) => 1.0 - i * 0.025)); // 1.0 → 0.275, strictly down
+                if (metric === "cpu")
+                    return series(Array.from({ length: 30 }, () => 20 + (Math.random() < 0 ? 1 : 0)));
+                return series(Array.from({ length: 30 }, () => 0.01)); // error_rate flat
+            },
+        };
+        reg.connectors.set("prom1", mock);
+        reg.sourceConfigs.set("prom1", { name: "prom1", type: "prometheus", url: "http://m", enabled: true });
+        const result = await getServiceHealthHandler(reg, { service: "payment-service" });
+        const data = JSON.parse(result.content[0].text);
+        const latAnom = (data.anomalies || []).find((a) => a.metric === "latency_p99");
+        assert.equal(latAnom, undefined, `latency dropping must not be an anomaly, got: ${JSON.stringify(latAnom)}`);
+    });
+});

package/dist/tools/list-services.d.ts

@@ -1,4 +1,5 @@
 import type { ConnectorRegistry } from "../connectors/registry.js";
+import { type RequestContext } from "../context.js";
 export declare const listServicesDefinition: {
     name: "list_services";
     description: string;
@@ -14,7 +15,7 @@ export declare const listServicesDefinition: {
 };
 export declare function listServicesHandler(registry: ConnectorRegistry, args: {
     filter?: string;
-}): Promise<{
+}, _ctx?: RequestContext): Promise<{
     content: {
         type: "text";
         text: string;

package/dist/tools/list-services.js

@@ -1,3 +1,4 @@
+import { defaultContext } from "../context.js";
 export const listServicesDefinition = {
     name: "list_services",
     description: "List all monitored services discovered across all connected backends. Returns service names, their data sources, and signal types (metrics/logs).",
@@ -11,7 +12,7 @@ export const listServicesDefinition = {
         },
     },
 };
-export async function listServicesHandler(registry, args) {
+export async function listServicesHandler(registry, args, _ctx = defaultContext()) {
     const connectors = registry.getAll();
     const allServices = [];
     for (const connector of connectors) {

package/dist/tools/list-sources.d.ts

@@ -1,4 +1,5 @@
 import type { ConnectorRegistry } from "../connectors/registry.js";
+import { type RequestContext } from "../context.js";
 export declare const listSourcesDefinition: {
     name: "list_sources";
     description: string;
@@ -7,7 +8,7 @@ export declare const listSourcesDefinition: {
         properties: {};
     };
 };
-export declare function listSourcesHandler(registry: ConnectorRegistry): Promise<{
+export declare function listSourcesHandler(registry: ConnectorRegistry, _ctx?: RequestContext): Promise<{
     content: {
         type: "text";
         text: string;

package/dist/tools/list-sources.js

@@ -1,3 +1,4 @@
+import { defaultContext } from "../context.js";
 export const listSourcesDefinition = {
     name: "list_sources",
     description: "List all configured observability backends and their connection status. Use this to discover what data sources are available.",
@@ -6,7 +7,7 @@ export const listSourcesDefinition = {
         properties: {},
     },
 };
-export async function listSourcesHandler(registry) {
+export async function listSourcesHandler(registry, _ctx = defaultContext()) {
     const healthResults = await registry.healthCheckAll();
     const connectors = registry.getAll();
     const sources = connectors.map((c) => ({

package/dist/tools/query-logs.d.ts

@@ -1,4 +1,5 @@
 import type { ConnectorRegistry } from "../connectors/registry.js";
+import { type RequestContext } from "../context.js";
 export declare const queryLogsDefinition: {
     name: "query_logs";
     description: string;
@@ -35,7 +36,7 @@ export declare function queryLogsHandler(registry: ConnectorRegistry, args: {
     duration?: string;
     level?: string;
     limit?: number;
-}): Promise<{
+}, _ctx?: RequestContext): Promise<{
     content: {
         type: "text";
         text: string;

package/dist/tools/query-logs.js

@@ -1,3 +1,4 @@
+import { defaultContext } from "../context.js";
 import { validateDuration, validateServiceName, errorResponse } from "./validation.js";
 export const queryLogsDefinition = {
     name: "query_logs",
@@ -29,7 +30,7 @@ export const queryLogsDefinition = {
         required: ["service"],
     },
 };
-export async function queryLogsHandler(registry, args) {
+export async function queryLogsHandler(registry, args, _ctx = defaultContext()) {
     const svcErr = validateServiceName(args.service);
     if (svcErr)
         return errorResponse(svcErr);

package/dist/tools/query-metrics.d.ts

@@ -1,4 +1,5 @@
 import type { ConnectorRegistry } from "../connectors/registry.js";
+import { type RequestContext } from "../context.js";
 export declare const queryMetricsDefinition: {
     name: "query_metrics";
     description: string;
@@ -35,7 +36,7 @@ export declare function queryMetricsHandler(registry: ConnectorRegistry, args: {
     duration?: string;
     source?: string;
     groupBy?: string;
-}): Promise<{
+}, _ctx?: RequestContext): Promise<{
     content: {
         type: "text";
         text: string;

package/dist/tools/query-metrics.js

@@ -1,3 +1,4 @@
+import { defaultContext } from "../context.js";
 import { validateDuration, validateMetricName, validateServiceName, errorResponse } from "./validation.js";
 export const queryMetricsDefinition = {
     name: "query_metrics",
@@ -29,7 +30,14 @@ export const queryMetricsDefinition = {
         required: ["service", "metric"],
     },
 };
-export async function queryMetricsHandler(registry, args) {
+export async function queryMetricsHandler(registry, args, _ctx = defaultContext()) {
+    // Coarse single-tenant source scoping: if the principal is restricted to a
+    // source allow-list, deny an explicit out-of-scope source.
+    if (_ctx.allowedSources &&
+        args.source &&
+        !_ctx.allowedSources.includes(args.source)) {
+        return errorResponse(`forbidden: source "${args.source}" is not in your allowed sources`);
+    }
     const svcErr = validateServiceName(args.service);
     if (svcErr)
         return errorResponse(svcErr);