Skip to content

net controller: preserve exact diagnostic problems in summaries #786

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

Closed
Patrick W. Healy (phealy) wants to merge 1 commit into from
Closed

net controller: preserve exact diagnostic problems in summaries #786

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
9 changes: 3 additions & 6 deletions cmd/unbounded-net-controller/cluster_status.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,7 @@ import (
"k8s.io/klog/v2"

"github.com/Azure/unbounded/internal/net/controller"
statuspkg "github.com/Azure/unbounded/internal/net/status"
statusv1alpha1 "github.com/Azure/unbounded/internal/net/status/v1alpha1"
"github.com/Azure/unbounded/internal/version"
)
Expand Down Expand Up @@ -821,12 +822,8 @@ func collectClusterProblems(status *ClusterStatusResponse) []StatusProblem {
}

if overview := status.NodeOverviews[node.NodeInfo.Name]; overview != nil {
if overview.RouteMismatch {
appendProblem("node", nodeName, "Route next-hop mismatches (expected vs present)")
}

if unhealthy := overview.PeerCount - overview.HealthyPeers; unhealthy > 0 {
appendProblem("node", nodeName, fmt.Sprintf("%d peers are not healthy", unhealthy))
for _, message := range statuspkg.OverviewDiagnosticMessages(*overview, node.NodeInfo.ProviderID) {
appendProblem("node", nodeName, message)
}

continue
Expand Down
1 change: 1 addition & 0 deletions cmd/unbounded-net-controller/node_overview_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -110,6 +110,7 @@ func TestClusterOverviewPreservesCountsAndEnrichment(t *testing.T) {
overview := statusv1alpha1.NodeStatusOverview{
NodeInfo: NodeInfo{Name: "node", SiteName: "site", WireGuard: &WireGuardStatusInfo{Interface: "wg0"}},
StatusSource: "ws", PeerCount: 20, HealthyPeers: 17, RouteCount: 30, RouteMismatch: true,
RouteMismatchCount: 2, UnhealthyPeerLinks: 3,
}
c.PatchOverview("node", overview)
snapshot := c.Get()
Expand Down
29 changes: 29 additions & 0 deletions cmd/unbounded-net-controller/overview_diagnostics_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -50,6 +50,35 @@ func TestOverviewDiagnosticMessagesMatchFullProblems(t *testing.T) {
t.Fatalf("IPIP warning applied outside Azure: %v", otherCloud)
}

metadata := statuspkg.OverviewMetadata(overview)
metadata.NodeInfo.ProviderID = node.NodeInfo.ProviderID
cluster := &ClusterStatusResponse{
Nodes: []*NodeStatusResponse{&metadata},
NodeOverviews: map[string]*statusv1alpha1.NodeStatusOverview{"node": &overview},
}

problems := collectClusterProblems(cluster)
if len(problems) != 1 || !slices.Equal(problems[0].Errors, fullProblems[0].Errors) {
t.Fatalf("overview problem pipeline=%+v, legacy=%+v", problems, fullProblems)
}

metadata.NodeInfo.ProviderID = "other://vm"

if problems = collectClusterProblems(cluster); len(problems) != 1 || len(problems[0].Errors) != 2 {
t.Fatalf("controller ignored enriched cloud identity: %+v", problems)
}

overview.RouteMismatch = false
overview.RouteMismatchCount = 0
overview.UnhealthyPeerLinks = 0

if problems = collectClusterProblems(cluster); len(problems) != 0 {
t.Fatalf("overview peer counts incorrectly replaced diagnostic link health: %+v", problems)
}

overview.RouteMismatchCount = 3
overview.UnhealthyPeerLinks = 2

overview.UsesIPIP = false
if noIPIP := statuspkg.OverviewDiagnosticMessages(overview, node.NodeInfo.ProviderID); len(noIPIP) != 2 {
t.Fatalf("Azure warning applied without IPIP: %v", noIPIP)
Expand Down