Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
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
41 changes: 18 additions & 23 deletions server/cmd/api/main.go
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ import (
"encoding/json"
"fmt"
"log/slog"
"net"
"net/http"
"net/url"
"os"
Expand Down Expand Up @@ -377,6 +378,7 @@ func main() {
metrics.NewChromeCollector(upstreamMgr),
metrics.NewGPUCollector(),
metrics.NewSystemCollector(),
metrics.NewResponseDrainCollector(scaletozero.ResponseDrainOutcomeCounts, scaletozero.ActiveResponseHolds, scaletozero.FailClosedResponseHolds),
}
if otlpMetrics != nil {
metricsCollectors = append(metricsCollectors, metrics.NewOTLPCollector(otlpMetrics))
Expand All @@ -387,29 +389,22 @@ func main() {
Handler: rMetrics,
}

go func() {
slogger.Info("http server starting", "addr", srv.Addr)
if err := srv.ListenAndServe(); err != nil && err != http.ErrServerClosed {
slogger.Error("http server failed", "err", err)
stop()
}
}()

go func() {
slogger.Info("devtools websocket proxy starting", "addr", srvDevtools.Addr)
if err := srvDevtools.ListenAndServe(); err != nil && err != http.ErrServerClosed {
slogger.Error("devtools websocket proxy failed", "err", err)
stop()
}
}()

go func() {
slogger.Info("chromedriver proxy starting", "addr", srvChromeDriver.Addr)
if err := srvChromeDriver.ListenAndServe(); err != nil && err != http.ErrServerClosed {
slogger.Error("chromedriver proxy failed", "err", err)
stop()
}
}()
serveHTTP := func(name string, server *http.Server) {
go func() {
slogger.Info(name+" starting", "addr", server.Addr)
listener, err := net.Listen("tcp", server.Addr)
if err == nil {
err = scaletozero.Serve(server, listener)
}
if err != nil && err != http.ErrServerClosed {
slogger.Error(name+" failed", "err", err)
stop()
}
}()
}
serveHTTP("http server", srv)
serveHTTP("devtools websocket proxy", srvDevtools)
serveHTTP("chromedriver proxy", srvChromeDriver)

go func() {
slogger.Info("metrics server starting", "addr", srvMetrics.Addr)
Expand Down
40 changes: 40 additions & 0 deletions server/lib/metrics/scaletozero.go
Original file line number Diff line number Diff line change
@@ -0,0 +1,40 @@
package metrics

import (
"context"
"sort"
)

type ResponseDrainSource func() map[string]uint64
type ResponseDrainGauge func() int64

type ResponseDrainCollector struct {
snapshot ResponseDrainSource
active ResponseDrainGauge
failClosed ResponseDrainGauge
}

func NewResponseDrainCollector(snapshot ResponseDrainSource, active, failClosed ResponseDrainGauge) *ResponseDrainCollector {
return &ResponseDrainCollector{snapshot: snapshot, active: active, failClosed: failClosed}
}

func (c *ResponseDrainCollector) Name() string { return "scale-to-zero response drain" }

func (c *ResponseDrainCollector) Collect(_ context.Context, w *Writer) error {
w.Metric("kernel_scale_to_zero_response_drain_total", "HTTP response drain events before scale-to-zero is re-enabled.", "counter")
counts := c.snapshot()
outcomes := make([]string, 0, len(counts))
for outcome := range counts {
outcomes = append(outcomes, outcome)
}
sort.Strings(outcomes)
for _, outcome := range outcomes {
w.Sample("kernel_scale_to_zero_response_drain_total", []Label{{Name: "outcome", Value: outcome}}, float64(counts[outcome]))
}

w.Metric("kernel_scale_to_zero_response_holds", "HTTP response scale-to-zero holds currently active.", "gauge")
w.Sample("kernel_scale_to_zero_response_holds", nil, float64(c.active()))
w.Metric("kernel_scale_to_zero_response_fail_closed_holds", "HTTP response holds awaiting terminal connection recovery.", "gauge")
w.Sample("kernel_scale_to_zero_response_fail_closed_holds", nil, float64(c.failClosed()))
return nil
}
33 changes: 33 additions & 0 deletions server/lib/metrics/scaletozero_test.go
Original file line number Diff line number Diff line change
@@ -0,0 +1,33 @@
package metrics

import (
"context"
"testing"

"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)

func TestResponseDrainCollector(t *testing.T) {
collector := NewResponseDrainCollector(func() map[string]uint64 {
return map[string]uint64{
"timeout": 2,
"drained": 7,
}
}, func() int64 { return 3 }, func() int64 { return 1 })
writer := &Writer{}

require.NoError(t, collector.Collect(context.Background(), writer))

assert.Equal(t, `# HELP kernel_scale_to_zero_response_drain_total HTTP response drain events before scale-to-zero is re-enabled.
# TYPE kernel_scale_to_zero_response_drain_total counter
kernel_scale_to_zero_response_drain_total{outcome="drained"} 7
kernel_scale_to_zero_response_drain_total{outcome="timeout"} 2
# HELP kernel_scale_to_zero_response_holds HTTP response scale-to-zero holds currently active.
# TYPE kernel_scale_to_zero_response_holds gauge
kernel_scale_to_zero_response_holds 3
# HELP kernel_scale_to_zero_response_fail_closed_holds HTTP response holds awaiting terminal connection recovery.
# TYPE kernel_scale_to_zero_response_fail_closed_holds gauge
kernel_scale_to_zero_response_fail_closed_holds 1
`, string(writer.Bytes()))
}
Loading
Loading