feat: expose current agent connections by type via prometheus (#14612)

This commit is contained in:
Ethan
2024-09-11 14:13:30 +10:00
committed by GitHub
parent 40688e40df
commit c8580a415a
5 changed files with 82 additions and 43 deletions
+52 -42
View File
@@ -2531,17 +2531,17 @@ func TestAgent_Metrics_SSH(t *testing.T) {
err = session.Shell()
require.NoError(t, err)
expected := []agentsdk.AgentMetric{
expected := []*proto.Stats_Metric{
{
Name: "agent_reconnecting_pty_connections_total",
Type: agentsdk.AgentMetricTypeCounter,
Type: proto.Stats_Metric_COUNTER,
Value: 0,
},
{
Name: "agent_sessions_total",
Type: agentsdk.AgentMetricTypeCounter,
Type: proto.Stats_Metric_COUNTER,
Value: 1,
Labels: []agentsdk.AgentMetricLabel{
Labels: []*proto.Stats_Metric_Label{
{
Name: "magic_type",
Value: "ssh",
@@ -2554,30 +2554,46 @@ func TestAgent_Metrics_SSH(t *testing.T) {
},
{
Name: "agent_ssh_server_failed_connections_total",
Type: agentsdk.AgentMetricTypeCounter,
Type: proto.Stats_Metric_COUNTER,
Value: 0,
},
{
Name: "agent_ssh_server_sftp_connections_total",
Type: agentsdk.AgentMetricTypeCounter,
Type: proto.Stats_Metric_COUNTER,
Value: 0,
},
{
Name: "agent_ssh_server_sftp_server_errors_total",
Type: agentsdk.AgentMetricTypeCounter,
Type: proto.Stats_Metric_COUNTER,
Value: 0,
},
{
Name: "coderd_agentstats_startup_script_seconds",
Type: agentsdk.AgentMetricTypeGauge,
Name: "coderd_agentstats_currently_reachable_peers",
Type: proto.Stats_Metric_GAUGE,
Value: 0,
Labels: []agentsdk.AgentMetricLabel{
Labels: []*proto.Stats_Metric_Label{
{
Name: "success",
Value: "true",
Name: "connection_type",
Value: "derp",
},
},
},
{
Name: "coderd_agentstats_currently_reachable_peers",
Type: proto.Stats_Metric_GAUGE,
Value: 1,
Labels: []*proto.Stats_Metric_Label{
{
Name: "connection_type",
Value: "p2p",
},
},
},
{
Name: "coderd_agentstats_startup_script_seconds",
Type: proto.Stats_Metric_GAUGE,
Value: 1,
},
}
var actual []*promgo.MetricFamily
@@ -2586,17 +2602,33 @@ func TestAgent_Metrics_SSH(t *testing.T) {
if err != nil {
return false
}
if len(expected) != len(actual) {
return false
count := 0
for _, m := range actual {
count += len(m.GetMetric())
}
return verifyCollectedMetrics(t, expected, actual)
return count == len(expected)
}, testutil.WaitLong, testutil.IntervalFast)
require.Len(t, actual, len(expected))
collected := verifyCollectedMetrics(t, expected, actual)
require.True(t, collected, "expected metrics were not collected")
i := 0
for _, mf := range actual {
for _, m := range mf.GetMetric() {
assert.Equal(t, expected[i].Name, mf.GetName())
assert.Equal(t, expected[i].Type.String(), mf.GetType().String())
// Value is max expected
if expected[i].Type == proto.Stats_Metric_GAUGE {
assert.GreaterOrEqualf(t, expected[i].Value, m.GetGauge().GetValue(), "expected %s to be greater than or equal to %f, got %f", expected[i].Name, expected[i].Value, m.GetGauge().GetValue())
} else if expected[i].Type == proto.Stats_Metric_COUNTER {
assert.GreaterOrEqualf(t, expected[i].Value, m.GetCounter().GetValue(), "expected %s to be greater than or equal to %f, got %f", expected[i].Name, expected[i].Value, m.GetCounter().GetValue())
}
for j, lbl := range expected[i].Labels {
assert.Equal(t, m.GetLabel()[j], &promgo.LabelPair{
Name: &lbl.Name,
Value: &lbl.Value,
})
}
i++
}
}
_ = stdin.Close()
err = session.Wait()
@@ -2828,28 +2860,6 @@ func TestAgent_ManageProcessPriority(t *testing.T) {
})
}
func verifyCollectedMetrics(t *testing.T, expected []agentsdk.AgentMetric, actual []*promgo.MetricFamily) bool {
t.Helper()
for i, e := range expected {
assert.Equal(t, e.Name, actual[i].GetName())
assert.Equal(t, string(e.Type), strings.ToLower(actual[i].GetType().String()))
for _, m := range actual[i].GetMetric() {
assert.Equal(t, e.Value, m.Counter.GetValue())
if len(m.GetLabel()) > 0 {
for j, lbl := range m.GetLabel() {
assert.Equal(t, e.Labels[j].Name, lbl.GetName())
assert.Equal(t, e.Labels[j].Value, lbl.GetValue())
}
}
m.GetLabel()
}
}
return true
}
type syncWriter struct {
mu sync.Mutex
w io.Writer