feat sicedar: added readiness prober

CHANGELOG.md

@@ -15,6 +15,7 @@ We use *breaking* word for marking changes that are not backward compatible (rel
 
 - [#1358](https://github.com/thanos-io/thanos/pull/1358) Added `part_size` configuration option for HTTP multipart requests minimum part size for S3 storage type
 - [#1363](https://github.com/thanos-io/thanos/pull/1363) Thanos Receive now exposes `thanos_receive_hashring_nodes` and `thanos_receive_hashring_tenants` metrics to monitor status of hash-rings
+- [#1395](https://github.com/thanos-io/thanos/pull/1395) Added `/-/ready` and `/-/healthy` endpoints to Thanos sidecar.
 
 ### Changed
 

cmd/thanos/compact.go

@@ -70,8 +70,7 @@ func (cs compactionSet) maxLevel() int {
 }
 
 func registerCompact(m map[string]setupFunc, app *kingpin.Application) {
-	comp := component.Compact
-	cmd := app.Command(comp.String(), "continuously compacts blocks in an object store bucket")
+	cmd := app.Command(component.Compact.String(), "continuously compacts blocks in an object store bucket")
 
 	haltOnError := cmd.Flag("debug.halt-on-error", "Halt the process if a critical compaction error is detected.").
 		Hidden().Default("true").Bool()
@@ -112,7 +111,7 @@ func registerCompact(m map[string]setupFunc, app *kingpin.Application) {
 	compactionConcurrency := cmd.Flag("compact.concurrency", "Number of goroutines to use when compacting groups.").
 		Default("1").Int()
 
-	m[comp.String()] = func(g *run.Group, logger log.Logger, reg *prometheus.Registry, tracer opentracing.Tracer, _ bool) error {
+	m[component.Compact.String()] = func(g *run.Group, logger log.Logger, reg *prometheus.Registry, tracer opentracing.Tracer, _ bool) error {
 		return runCompact(g, logger, reg,
 			*httpAddr,
 			*dataDir,
@@ -127,7 +126,7 @@ func registerCompact(m map[string]setupFunc, app *kingpin.Application) {
 				compact.ResolutionLevel5m:  time.Duration(*retention5m),
 				compact.ResolutionLevel1h:  time.Duration(*retention1h),
 			},
-			comp,
+			component.Compact,
 			*disableDownsampling,
 			*maxCompactionLevel,
 			*blockSyncConcurrency,
@@ -170,9 +169,9 @@ func runCompact(
 
 	downsampleMetrics := newDownsampleMetrics(reg)
 
-	readinessProber := prober.NewProber(component, logger, prometheus.WrapRegistererWithPrefix("thanos_", reg))
+	statusProber := prober.NewProber(component, logger, prometheus.WrapRegistererWithPrefix("thanos_", reg))
 	// Initiate default HTTP listener providing metrics endpoint and readiness/liveness probes.
-	if err := defaultHTTPListener(g, logger, reg, httpBindAddr, readinessProber); err != nil {
+	if err := defaultHTTPListener(g, logger, reg, httpBindAddr, statusProber); err != nil {
 		return errors.Wrap(err, "create readiness prober")
 	}
 
@@ -327,7 +326,7 @@ func runCompact(
 	})
 
 	level.Info(logger).Log("msg", "starting compact node")
-	readinessProber.SetReady()
+	statusProber.SetReady()
 	return nil
 }
 

cmd/thanos/main.go

@@ -70,7 +70,7 @@ func main() {
 	tracingConfig := regCommonTracingFlags(app)
 
 	cmds := map[string]setupFunc{}
-	registerSidecar(cmds, app, "sidecar")
+	registerSidecar(cmds, app)
 	registerStore(cmds, app, "store")
 	registerQuery(cmds, app, "query")
 	registerRule(cmds, app, "rule")

cmd/thanos/sidecar.go

@@ -11,28 +11,29 @@ import (
 	"github.com/go-kit/kit/log"
 	"github.com/go-kit/kit/log/level"
 	"github.com/oklog/run"
-	opentracing "github.com/opentracing/opentracing-go"
+	"github.com/opentracing/opentracing-go"
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/common/model"
 	"github.com/prometheus/tsdb/labels"
 	"github.com/thanos-io/thanos/pkg/block/metadata"
 	"github.com/thanos-io/thanos/pkg/component"
 	"github.com/thanos-io/thanos/pkg/objstore/client"
+	"github.com/thanos-io/thanos/pkg/prober"
 	"github.com/thanos-io/thanos/pkg/promclient"
 	"github.com/thanos-io/thanos/pkg/reloader"
 	"github.com/thanos-io/thanos/pkg/runutil"
 	"github.com/thanos-io/thanos/pkg/shipper"
 	"github.com/thanos-io/thanos/pkg/store"
 	"github.com/thanos-io/thanos/pkg/store/storepb"
 	"google.golang.org/grpc"
-	kingpin "gopkg.in/alecthomas/kingpin.v2"
+	"gopkg.in/alecthomas/kingpin.v2"
 )
 
 const waitForExternalLabelsTimeout = 10 * time.Minute
 
-func registerSidecar(m map[string]setupFunc, app *kingpin.Application, name string) {
-	cmd := app.Command(name, "sidecar for Prometheus server")
+func registerSidecar(m map[string]setupFunc, app *kingpin.Application) {
+	cmd := app.Command(component.Sidecar.String(), "sidecar for Prometheus server")
 
 	grpcBindAddr, httpBindAddr, cert, key, clientCA := regCommonServerFlags(cmd)
 
@@ -54,7 +55,7 @@ func registerSidecar(m map[string]setupFunc, app *kingpin.Application, name stri
 
 	uploadCompacted := cmd.Flag("shipper.upload-compacted", "[Experimental] If true sidecar will try to upload compacted blocks as well. Useful for migration purposes. Works only if compaction is disabled on Prometheus.").Default("false").Hidden().Bool()
 
-	m[name] = func(g *run.Group, logger log.Logger, reg *prometheus.Registry, tracer opentracing.Tracer, _ bool) error {
+	m[component.Sidecar.String()] = func(g *run.Group, logger log.Logger, reg *prometheus.Registry, tracer opentracing.Tracer, _ bool) error {
 		rl := reloader.New(
 			log.With(logger, "component", "reloader"),
 			reloader.ReloadURLFromBase(*promURL),
@@ -77,6 +78,7 @@ func registerSidecar(m map[string]setupFunc, app *kingpin.Application, name stri
 			objStoreConfig,
 			rl,
 			*uploadCompacted,
+			component.Sidecar,
 		)
 	}
 }
@@ -96,6 +98,7 @@ func runSidecar(
 	objStoreConfig *pathOrContent,
 	reloader *reloader.Reloader,
 	uploadCompacted bool,
+	comp component.Component,
 ) error {
 	var m = &promMetadata{
 		promURL: promURL,
@@ -117,6 +120,12 @@ func runSidecar(
 		uploads = false
 	}
 
+	statusProber := prober.NewProber(comp, logger, prometheus.WrapRegistererWithPrefix("thanos_", reg))
+	// Initiate default HTTP listener providing metrics endpoint and readiness/liveness probes.
+	if err := defaultHTTPListener(g, logger, reg, httpBindAddr, statusProber); err != nil {
+		return errors.Wrap(err, "create readiness prober")
+	}
+
 	// Setup all the concurrent groups.
 	{
 		promUp := prometheus.NewGauge(prometheus.GaugeOpts{
@@ -148,6 +157,7 @@ func runSidecar(
 						"err", err,
 					)
 					promUp.Set(0)
+					statusProber.SetNotReady(err)
 					return err
 				}
 
@@ -156,6 +166,7 @@ func runSidecar(
 					"external_labels", m.Labels().String(),
 				)
 				promUp.Set(1)
+				statusProber.SetReady()
 				lastHeartbeat.Set(float64(time.Now().UnixNano()) / 1e9)
 				return nil
 			})
@@ -195,9 +206,7 @@ func runSidecar(
 			cancel()
 		})
 	}
-	if err := metricHTTPListenGroup(g, logger, reg, httpBindAddr); err != nil {
-		return err
-	}
+
 	{
 		l, err := net.Listen("tcp", grpcBindAddr)
 		if err != nil {

pkg/prober/prober.go

@@ -22,6 +22,19 @@ const (
 )
 
 // Prober represents health and readiness status of given component.
+//
+// From Kubernetes documentation https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-probes/ :
+//
+//   liveness: Many applications running for long periods of time eventually transition to broken states,
+//   (healthy) and cannot recover except by being restarted.
+//             Kubernetes provides liveness probes to detect and remedy such situations.
+//
+//   readiness: Sometimes, applications are temporarily unable to serve traffic.
+//   (ready)    For example, an application might need to load large data or configuration files during startup,
+//              or depend on external services after startup. In such cases, you don’t want to kill the application,
+//              but you don’t want to send it requests either. Kubernetes provides readiness probes to detect
+//              and mitigate these situations. A pod with containers reporting that they are not ready
+//              does not receive traffic through Kubernetes Services.
 type Prober struct {
 	logger             log.Logger
 	component          component.Component
@@ -36,7 +49,6 @@ type Prober struct {
 // NewProber returns Prober representing readiness and healthiness of given component.
 func NewProber(component component.Component, logger log.Logger, reg prometheus.Registerer) *Prober {
 	initialErr := fmt.Errorf(initialErrorFmt, component)
-
 	p := &Prober{
 		component:   component,
 		logger:      logger,
@@ -76,7 +88,7 @@ func (p *Prober) writeResponse(w http.ResponseWriter, probeFn func() error, prob
 		http.Error(w, fmt.Sprintf("thanos %v is not %v. Reason: %v", p.component, probeType, err), probeErrorHTTPStatus)
 		return
 	}
-	if _, err := io.WriteString(w, fmt.Sprintf("thanos %v is %v", p.component, probeType)); err == nil {
+	if _, err := io.WriteString(w, fmt.Sprintf("thanos %v is %v", p.component, probeType)); err != nil {
 		level.Error(p.logger).Log("msg", "failed to write probe response", "probe type", probeType, "err", err)
 	}
 }

tutorials/kubernetes-demo/manifests/prometheus-ha-sidecar-lts.yaml

@@ -145,6 +145,14 @@ spec:
             containerPort: 10902
           - name: grpc
             containerPort: 10901
+        livenessProbe:
+            httpGet:
+              port: 10902
+              path: /-/healthy
+        readinessProbe:
+          httpGet:
+            port: 10902
+            path: /-/ready
         volumeMounts:
           - name: prometheus
             mountPath: /var/prometheus

tutorials/kubernetes-demo/manifests/prometheus-ha-sidecar.yaml

@@ -129,6 +129,14 @@ spec:
             containerPort: 10902
           - name: grpc
             containerPort: 10901
+        livenessProbe:
+            httpGet:
+              port: 10902
+              path: /-/healthy
+        readinessProbe:
+          httpGet:
+            port: 10902
+            path: /-/ready
         volumeMounts:
           - name: prometheus
             mountPath: /var/prometheus