New MCP server that queries live Prometheus and Alertmanager HTTP APIs with 8 tools: list_alerts, get_alert, search_metrics, get_metric_metadata, query (PromQL), list_targets, list_silences, and create_silence. Extends the MCP core with ModeCustom and NewGenericServer for servers that don't require a database. Includes CLI with direct commands (alerts, query, targets, metrics), NixOS module, and comprehensive httptest-based tests. Bumps existing binaries to 0.2.1 due to shared internal/mcp change. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
379 lines
10 KiB
Go
379 lines
10 KiB
Go
package monitoring
|
|
|
|
import (
|
|
"context"
|
|
"encoding/json"
|
|
"io"
|
|
"log"
|
|
"net/http"
|
|
"net/http/httptest"
|
|
"strings"
|
|
"testing"
|
|
|
|
"git.t-juice.club/torjus/labmcp/internal/mcp"
|
|
)
|
|
|
|
// setupTestServer creates a test MCP server with monitoring handlers backed by test HTTP servers.
|
|
func setupTestServer(t *testing.T, promHandler, amHandler http.HandlerFunc) (*mcp.Server, func()) {
|
|
t.Helper()
|
|
|
|
promSrv := httptest.NewServer(promHandler)
|
|
amSrv := httptest.NewServer(amHandler)
|
|
|
|
logger := log.New(io.Discard, "", 0)
|
|
config := mcp.DefaultMonitoringConfig()
|
|
server := mcp.NewGenericServer(logger, config)
|
|
|
|
prom := NewPrometheusClient(promSrv.URL)
|
|
am := NewAlertmanagerClient(amSrv.URL)
|
|
RegisterHandlers(server, prom, am)
|
|
|
|
cleanup := func() {
|
|
promSrv.Close()
|
|
amSrv.Close()
|
|
}
|
|
|
|
return server, cleanup
|
|
}
|
|
|
|
func TestHandler_ListAlerts(t *testing.T) {
|
|
server, cleanup := setupTestServer(t,
|
|
nil,
|
|
func(w http.ResponseWriter, r *http.Request) {
|
|
w.Header().Set("Content-Type", "application/json")
|
|
_, _ = w.Write([]byte(`[
|
|
{
|
|
"annotations": {"summary": "Node is down"},
|
|
"endsAt": "2024-01-01T01:00:00Z",
|
|
"fingerprint": "fp1",
|
|
"receivers": [{"name": "default"}],
|
|
"startsAt": "2024-01-01T00:00:00Z",
|
|
"status": {"inhibitedBy": [], "silencedBy": [], "state": "active"},
|
|
"updatedAt": "2024-01-01T00:00:00Z",
|
|
"generatorURL": "",
|
|
"labels": {"alertname": "NodeDown", "severity": "critical"}
|
|
}
|
|
]`))
|
|
},
|
|
)
|
|
defer cleanup()
|
|
|
|
result := callTool(t, server, "list_alerts", map[string]interface{}{})
|
|
if result.IsError {
|
|
t.Fatalf("unexpected error: %s", result.Content[0].Text)
|
|
}
|
|
if !strings.Contains(result.Content[0].Text, "NodeDown") {
|
|
t.Errorf("expected output to contain 'NodeDown', got: %s", result.Content[0].Text)
|
|
}
|
|
if !strings.Contains(result.Content[0].Text, "1 alert") {
|
|
t.Errorf("expected output to contain '1 alert', got: %s", result.Content[0].Text)
|
|
}
|
|
}
|
|
|
|
func TestHandler_GetAlert(t *testing.T) {
|
|
server, cleanup := setupTestServer(t,
|
|
nil,
|
|
func(w http.ResponseWriter, r *http.Request) {
|
|
w.Header().Set("Content-Type", "application/json")
|
|
_, _ = w.Write([]byte(`[
|
|
{
|
|
"annotations": {"summary": "Found it"},
|
|
"endsAt": "2024-01-01T01:00:00Z",
|
|
"fingerprint": "target-fp",
|
|
"receivers": [{"name": "default"}],
|
|
"startsAt": "2024-01-01T00:00:00Z",
|
|
"status": {"inhibitedBy": [], "silencedBy": [], "state": "active"},
|
|
"updatedAt": "2024-01-01T00:00:00Z",
|
|
"generatorURL": "",
|
|
"labels": {"alertname": "TestAlert", "severity": "warning"}
|
|
},
|
|
{
|
|
"annotations": {},
|
|
"endsAt": "2024-01-01T01:00:00Z",
|
|
"fingerprint": "other-fp",
|
|
"receivers": [{"name": "default"}],
|
|
"startsAt": "2024-01-01T00:00:00Z",
|
|
"status": {"inhibitedBy": [], "silencedBy": [], "state": "active"},
|
|
"updatedAt": "2024-01-01T00:00:00Z",
|
|
"generatorURL": "",
|
|
"labels": {"alertname": "OtherAlert", "severity": "info"}
|
|
}
|
|
]`))
|
|
},
|
|
)
|
|
defer cleanup()
|
|
|
|
result := callTool(t, server, "get_alert", map[string]interface{}{
|
|
"fingerprint": "target-fp",
|
|
})
|
|
if result.IsError {
|
|
t.Fatalf("unexpected error: %s", result.Content[0].Text)
|
|
}
|
|
if !strings.Contains(result.Content[0].Text, "TestAlert") {
|
|
t.Errorf("expected output to contain 'TestAlert', got: %s", result.Content[0].Text)
|
|
}
|
|
}
|
|
|
|
func TestHandler_GetAlertNotFound(t *testing.T) {
|
|
server, cleanup := setupTestServer(t,
|
|
nil,
|
|
func(w http.ResponseWriter, r *http.Request) {
|
|
w.Header().Set("Content-Type", "application/json")
|
|
_, _ = w.Write([]byte(`[]`))
|
|
},
|
|
)
|
|
defer cleanup()
|
|
|
|
result := callTool(t, server, "get_alert", map[string]interface{}{
|
|
"fingerprint": "nonexistent",
|
|
})
|
|
if !result.IsError {
|
|
t.Error("expected error result for nonexistent fingerprint")
|
|
}
|
|
}
|
|
|
|
func TestHandler_Query(t *testing.T) {
|
|
server, cleanup := setupTestServer(t,
|
|
func(w http.ResponseWriter, r *http.Request) {
|
|
if r.URL.Path != "/api/v1/query" {
|
|
http.NotFound(w, r)
|
|
return
|
|
}
|
|
w.Header().Set("Content-Type", "application/json")
|
|
_, _ = w.Write([]byte(`{
|
|
"status": "success",
|
|
"data": {
|
|
"resultType": "vector",
|
|
"result": [
|
|
{
|
|
"metric": {"__name__": "up", "job": "node"},
|
|
"value": [1234567890, "1"]
|
|
}
|
|
]
|
|
}
|
|
}`))
|
|
},
|
|
nil,
|
|
)
|
|
defer cleanup()
|
|
|
|
result := callTool(t, server, "query", map[string]interface{}{
|
|
"promql": "up",
|
|
})
|
|
if result.IsError {
|
|
t.Fatalf("unexpected error: %s", result.Content[0].Text)
|
|
}
|
|
if !strings.Contains(result.Content[0].Text, "node") {
|
|
t.Errorf("expected output to contain 'node', got: %s", result.Content[0].Text)
|
|
}
|
|
}
|
|
|
|
func TestHandler_ListTargets(t *testing.T) {
|
|
server, cleanup := setupTestServer(t,
|
|
func(w http.ResponseWriter, r *http.Request) {
|
|
if r.URL.Path != "/api/v1/targets" {
|
|
http.NotFound(w, r)
|
|
return
|
|
}
|
|
w.Header().Set("Content-Type", "application/json")
|
|
_, _ = w.Write([]byte(`{
|
|
"status": "success",
|
|
"data": {
|
|
"activeTargets": [
|
|
{
|
|
"labels": {"instance": "localhost:9090", "job": "prometheus"},
|
|
"scrapePool": "prometheus",
|
|
"scrapeUrl": "http://localhost:9090/metrics",
|
|
"globalUrl": "http://localhost:9090/metrics",
|
|
"lastError": "",
|
|
"lastScrape": "2024-01-01T00:00:00Z",
|
|
"lastScrapeDuration": 0.015,
|
|
"health": "up",
|
|
"scrapeInterval": "15s",
|
|
"scrapeTimeout": "10s"
|
|
}
|
|
],
|
|
"droppedTargets": []
|
|
}
|
|
}`))
|
|
},
|
|
nil,
|
|
)
|
|
defer cleanup()
|
|
|
|
result := callTool(t, server, "list_targets", map[string]interface{}{})
|
|
if result.IsError {
|
|
t.Fatalf("unexpected error: %s", result.Content[0].Text)
|
|
}
|
|
if !strings.Contains(result.Content[0].Text, "prometheus") {
|
|
t.Errorf("expected output to contain 'prometheus', got: %s", result.Content[0].Text)
|
|
}
|
|
}
|
|
|
|
func TestHandler_SearchMetrics(t *testing.T) {
|
|
server, cleanup := setupTestServer(t,
|
|
func(w http.ResponseWriter, r *http.Request) {
|
|
w.Header().Set("Content-Type", "application/json")
|
|
switch r.URL.Path {
|
|
case "/api/v1/label/__name__/values":
|
|
_, _ = w.Write([]byte(`{
|
|
"status": "success",
|
|
"data": ["node_cpu_seconds_total", "node_memory_MemTotal_bytes", "up"]
|
|
}`))
|
|
case "/api/v1/metadata":
|
|
_, _ = w.Write([]byte(`{
|
|
"status": "success",
|
|
"data": {
|
|
"node_cpu_seconds_total": [{"type": "counter", "help": "CPU time", "unit": ""}],
|
|
"node_memory_MemTotal_bytes": [{"type": "gauge", "help": "Total memory", "unit": "bytes"}]
|
|
}
|
|
}`))
|
|
default:
|
|
http.NotFound(w, r)
|
|
}
|
|
},
|
|
nil,
|
|
)
|
|
defer cleanup()
|
|
|
|
result := callTool(t, server, "search_metrics", map[string]interface{}{
|
|
"query": "node",
|
|
})
|
|
if result.IsError {
|
|
t.Fatalf("unexpected error: %s", result.Content[0].Text)
|
|
}
|
|
if !strings.Contains(result.Content[0].Text, "node_cpu") {
|
|
t.Errorf("expected output to contain 'node_cpu', got: %s", result.Content[0].Text)
|
|
}
|
|
// "up" should be filtered out since it doesn't match "node"
|
|
if strings.Contains(result.Content[0].Text, "| up |") {
|
|
t.Errorf("expected 'up' to be filtered out, got: %s", result.Content[0].Text)
|
|
}
|
|
}
|
|
|
|
func TestHandler_ListSilences(t *testing.T) {
|
|
server, cleanup := setupTestServer(t,
|
|
nil,
|
|
func(w http.ResponseWriter, r *http.Request) {
|
|
if r.URL.Path != "/api/v2/silences" {
|
|
http.NotFound(w, r)
|
|
return
|
|
}
|
|
w.Header().Set("Content-Type", "application/json")
|
|
_, _ = w.Write([]byte(`[
|
|
{
|
|
"id": "s1",
|
|
"matchers": [{"name": "alertname", "value": "Test", "isRegex": false}],
|
|
"startsAt": "2024-01-01T00:00:00Z",
|
|
"endsAt": "2024-01-01T02:00:00Z",
|
|
"createdBy": "admin",
|
|
"comment": "Testing",
|
|
"status": {"state": "active"}
|
|
},
|
|
{
|
|
"id": "s2",
|
|
"matchers": [{"name": "job", "value": "node", "isRegex": false}],
|
|
"startsAt": "2023-01-01T00:00:00Z",
|
|
"endsAt": "2023-01-01T02:00:00Z",
|
|
"createdBy": "admin",
|
|
"comment": "Old",
|
|
"status": {"state": "expired"}
|
|
}
|
|
]`))
|
|
},
|
|
)
|
|
defer cleanup()
|
|
|
|
result := callTool(t, server, "list_silences", map[string]interface{}{})
|
|
if result.IsError {
|
|
t.Fatalf("unexpected error: %s", result.Content[0].Text)
|
|
}
|
|
// Should show active silence but filter out expired
|
|
if !strings.Contains(result.Content[0].Text, "s1") {
|
|
t.Errorf("expected active silence s1 in output, got: %s", result.Content[0].Text)
|
|
}
|
|
if strings.Contains(result.Content[0].Text, "s2") {
|
|
t.Errorf("expected expired silence s2 to be filtered out, got: %s", result.Content[0].Text)
|
|
}
|
|
}
|
|
|
|
func TestHandler_ToolCount(t *testing.T) {
|
|
server, cleanup := setupTestServer(t,
|
|
func(w http.ResponseWriter, r *http.Request) {},
|
|
func(w http.ResponseWriter, r *http.Request) {},
|
|
)
|
|
defer cleanup()
|
|
|
|
// Send a tools/list request
|
|
req := &mcp.Request{
|
|
JSONRPC: "2.0",
|
|
ID: 1,
|
|
Method: "tools/list",
|
|
}
|
|
resp := server.HandleRequest(context.Background(), req)
|
|
if resp == nil {
|
|
t.Fatal("expected response, got nil")
|
|
}
|
|
if resp.Error != nil {
|
|
t.Fatalf("unexpected error: %s", resp.Error.Message)
|
|
}
|
|
|
|
resultJSON, err := json.Marshal(resp.Result)
|
|
if err != nil {
|
|
t.Fatalf("failed to marshal result: %v", err)
|
|
}
|
|
|
|
var listResult mcp.ListToolsResult
|
|
if err := json.Unmarshal(resultJSON, &listResult); err != nil {
|
|
t.Fatalf("failed to unmarshal result: %v", err)
|
|
}
|
|
|
|
if len(listResult.Tools) != 8 {
|
|
t.Errorf("expected 8 tools, got %d", len(listResult.Tools))
|
|
for _, tool := range listResult.Tools {
|
|
t.Logf(" tool: %s", tool.Name)
|
|
}
|
|
}
|
|
}
|
|
|
|
// callTool is a test helper that calls a tool through the MCP server.
|
|
func callTool(t *testing.T, server *mcp.Server, name string, args map[string]interface{}) mcp.CallToolResult {
|
|
t.Helper()
|
|
|
|
params := mcp.CallToolParams{
|
|
Name: name,
|
|
Arguments: args,
|
|
}
|
|
paramsJSON, err := json.Marshal(params)
|
|
if err != nil {
|
|
t.Fatalf("failed to marshal params: %v", err)
|
|
}
|
|
|
|
req := &mcp.Request{
|
|
JSONRPC: "2.0",
|
|
ID: 1,
|
|
Method: "tools/call",
|
|
Params: paramsJSON,
|
|
}
|
|
|
|
resp := server.HandleRequest(context.Background(), req)
|
|
if resp == nil {
|
|
t.Fatal("expected response, got nil")
|
|
}
|
|
if resp.Error != nil {
|
|
t.Fatalf("JSON-RPC error: %s", resp.Error.Message)
|
|
}
|
|
|
|
resultJSON, err := json.Marshal(resp.Result)
|
|
if err != nil {
|
|
t.Fatalf("failed to marshal result: %v", err)
|
|
}
|
|
|
|
var result mcp.CallToolResult
|
|
if err := json.Unmarshal(resultJSON, &result); err != nil {
|
|
t.Fatalf("failed to unmarshal result: %v", err)
|
|
}
|
|
|
|
return result
|
|
}
|