Skip to content

feat: add basic support for multicluster env in MCP #71

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Nov 10, 2025
Merged

feat: add basic support for multicluster env in MCP #71

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b5c86e1
feat: add basic support for multicluster env in MCP
tremes Nov 5, 2025
2cc7a08
fix after review
tremes Nov 5, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 17 additions & 1 deletion pkg/mcp/incidents_tool.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -222,6 +222,8 @@ func (i *IncidentTool) transformPromValueToIncident(dataVec prom.RangeVector, qR
healthyVal := processor.HealthValue(lastSample.Value)
groupId := string(v.Metric["group_id"])
component := string(v.Metric["component"])
clusterName := string(v.Metric["cluster"])
clusterID := string(v.Metric["clusterID"])

if existingInc, ok := incidents[groupId]; ok {
existingInc.ComponentsSet[component] = struct{}{}
Expand Down Expand Up @@ -250,7 +252,9 @@ func (i *IncidentTool) transformPromValueToIncident(dataVec prom.RangeVector, qR
incidents[existingInc.GroupId] = existingInc
} else {
incident := Incident{
GroupId: string(groupId),
Cluster: clusterName,
ClusterID: clusterID,
GroupId: groupId,
Severity: healthyVal.String(),
StartTime: formatToRFC3339(startTime),
EndTime: formatToRFC3339(endTime),
Expand Down Expand Up @@ -326,6 +330,14 @@ func getAlertDataForIncidents(ctx context.Context, incidents map[string]Incident
for _, alertInIncident := range inc.Alerts {
subsetMatcher := common.LabelsSubsetMatcher{Labels: alertInIncident}
for _, firingAlert := range alerts {
// check for multicluster/ACM environment
if inc.ClusterID != "" {
clusterIDMatch := string(firingAlert["clusterID"]) == inc.ClusterID
// if the alert cluster ID does not match incident cluster ID, skip
if !clusterIDMatch {
continue
}
}
match, _ := subsetMatcher.Matches(firingAlert)
if match {

Expand Down Expand Up @@ -381,11 +393,15 @@ func cleanupLabels(m model.LabelSet) model.LabelSet {
updatedLS := m.Clone()
updatedLS["status"] = updatedLS["alertstate"]
updatedLS["name"] = updatedLS["alertname"]
if clusterID := updatedLS["clusterID"]; clusterID != "" {
updatedLS["cluster_id"] = clusterID
}
delete(updatedLS, "__name__")
delete(updatedLS, "prometheus")
delete(updatedLS, "alertstate")
delete(updatedLS, "alertname")
delete(updatedLS, "pod")
delete(updatedLS, "clusterID")
return updatedLS
}

Expand Down
158 changes: 158 additions & 0 deletions pkg/mcp/incidents_tool_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -824,6 +824,164 @@ func TestGetAlertDataForIncidents(t *testing.T) {
},
},
},
{
name: "Alerts are correctly matched in multicluster environment",
promLoader: func() prom.Loader {
mocked := mocks.NewMockPrometheusLoader(ctrl)
mocked.EXPECT().LoadVectorRange(gomock.Any(), `ALERTS{alertstate!="pending"}`, gomock.Any(), gomock.Any(), gomock.Any()).Return(prom.RangeVector{
{
Metric: model.LabelSet{
"alertname": "Alert1",
"namespace": "foo",
"pod": "red",
"alertstate": "firing",
"severity": "warning",
"clusterID": "1111",
},
Samples: []model.SamplePair{
{
Value: 1,
Timestamp: model.Now().Add(-20 * time.Minute),
},
{
Value: 1,
Timestamp: model.Now().Add(-1 * time.Minute),
},
},
},
{
Metric: model.LabelSet{
"alertname": "Alert1",
"namespace": "foo",
"pod": "red",
"alertstate": "firing",
"severity": "warning",
"clusterID": "2222",
},
Samples: []model.SamplePair{
{
Value: 1,
Timestamp: model.Now().Add(-20 * time.Minute),
},
{
Value: 1,
Timestamp: model.Now().Add(-1 * time.Minute),
},
},
},
{
Metric: model.LabelSet{
"alertname": "Alert1",
"namespace": "bar",
"pod": "blue",
"alertstate": "firing",
"severity": "critical",
"clusterID": "1111",
},
Samples: []model.SamplePair{
{
Value: 1,
Timestamp: model.Now().Add(-20 * time.Minute),
},
{
Value: 1,
Timestamp: model.Now().Add(-1 * time.Minute),
},
},
},
{
Metric: model.LabelSet{
"alertname": "Alert1",
"namespace": "bar",
"pod": "blue",
"alertstate": "firing",
"severity": "critical",
"clusterID": "2222",
},
Samples: []model.SamplePair{
{
Value: 1,
Timestamp: model.Now().Add(-20 * time.Minute),
},
{
Value: 1,
Timestamp: model.Now().Add(-1 * time.Minute),
},
},
},
}, nil)
return mocked
}(),
incidentsMap: map[string]Incident{
"1": {
GroupId: "1",
ClusterID: "1111",
Alerts: []model.LabelSet{
{"alertname": "Alert1", "namespace": "foo", "severity": "warning"},
{"alertname": "Alert1", "namespace": "bar", "severity": "critical"},
},
},
"2": {
GroupId: "2",
ClusterID: "2222",
Alerts: []model.LabelSet{
{"alertname": "Alert1", "namespace": "foo", "severity": "warning"},
{"alertname": "Alert1", "namespace": "bar", "severity": "critical"},
},
},
},
silencedAlerts: []models.Alert{},
expectedIncidents: []Incident{
{
GroupId: "1",
ClusterID: "1111",
Alerts: []model.LabelSet{
{
"name": "Alert1",
"namespace": "foo",
"status": "firing",
"silenced": "false",
"severity": "warning",
"cluster_id": "1111",
"start_time": model.LabelValue(model.Now().Add(-20 * time.Minute).Time().Format(time.RFC3339)),
},
{
"name": "Alert1",
"namespace": "bar",
"status": "firing",
"silenced": "false",
"severity": "critical",
"cluster_id": "1111",
"start_time": model.LabelValue(model.Now().Add(-20 * time.Minute).Time().Format(time.RFC3339)),
},
},
},
{
GroupId: "2",
ClusterID: "2222",
Alerts: []model.LabelSet{
{
"name": "Alert1",
"namespace": "foo",
"status": "firing",
"silenced": "false",
"severity": "warning",
"cluster_id": "2222",
"start_time": model.LabelValue(model.Now().Add(-20 * time.Minute).Time().Format(time.RFC3339)),
},
{
"name": "Alert1",
"namespace": "bar",
"status": "firing",
"silenced": "false",
"severity": "critical",
"cluster_id": "2222",
"start_time": model.LabelValue(model.Now().Add(-20 * time.Minute).Time().Format(time.RFC3339)),
},
},
},
},
},
}

for _, tt := range tests {
Expand Down
2 changes: 2 additions & 0 deletions pkg/mcp/types.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,8 @@ type Incident struct {
StartTime string `json:"start_time"`
Status string `json:"status"`
EndTime string `json:"end_time"`
Cluster string `json:"cluster,omitempty"`
ClusterID string `json:"cluster_id,omitempty"`

URL string `json:"url_details"`
Alerts []model.LabelSet `json:"alerts"`
Expand Down