rebased

Signed-off-by: Krasi Georgiev <8903888+krasi-georgiev@users.noreply.github.com>

CHANGELOG.md

@@ -25,6 +25,8 @@ We use *breaking* word for marking changes that are not backward compatible (rel
 - [#1758](https://github.com/thanos-io/thanos/pull/1758) Bucket: `thanos bucket web` now supports `--web.external-prefix` for proxying on a subpath.
 - [#1770](https://github.com/thanos-io/thanos/pull/1770) Bucket: Add `--web.prefix-header` flags to allow for bucket UI to be accessible behind a reverse proxy.
 - [#1668](https://github.com/thanos-io/thanos/pull/1668) Receiver: Added TLS options for both server and client remote write.
+- [#1672](https://github.com/thanos-io/thanos/pull/1672) All components: Client and server certificates are auto reloaded when changed on disk.
+- [#1672](https://github.com/thanos-io/thanos/pull/1672) Add a new `--grpc-server-max-connection-age` CLI option which controls how often to re-do the tls handshake and re-read the certificates. This in reality controls the keep alive so when the connection is closed it requires a new tls handshake.
 
 ### Fixed
 

cmd/thanos/flags.go

@@ -3,6 +3,7 @@ package main
 import (
 	"fmt"
 	"strings"
+	"time"
 
 	"github.com/thanos-io/thanos/pkg/extflag"
 
@@ -23,6 +24,7 @@ func regGRPCFlags(cmd *kingpin.CmdClause) (
 	grpcTLSSrvCert *string,
 	grpcTLSSrvKey *string,
 	grpcTLSSrvClientCA *string,
+	grpcMaxConnectionAge *time.Duration,
 ) {
 	grpcBindAddr = cmd.Flag("grpc-address", "Listen ip:port address for gRPC endpoints (StoreAPI). Make sure this address is routable from other components.").
 		Default("0.0.0.0:10901").String()
@@ -31,12 +33,14 @@ func regGRPCFlags(cmd *kingpin.CmdClause) (
 	grpcTLSSrvCert = cmd.Flag("grpc-server-tls-cert", "TLS Certificate for gRPC server, leave blank to disable TLS").Default("").String()
 	grpcTLSSrvKey = cmd.Flag("grpc-server-tls-key", "TLS Key for the gRPC server, leave blank to disable TLS").Default("").String()
 	grpcTLSSrvClientCA = cmd.Flag("grpc-server-tls-client-ca", "TLS CA to verify clients against. If no client CA is specified, there is no client verification on server side. (tls.NoClientCert)").Default("").String()
+	grpcMaxConnectionAge = cmd.Flag("grpc-server-max-connection-age", "The grpc server max connection age. This controls how often to re-read the tls certificates and redo the TLS handshake ").Default("1m").Duration()
 
 	return grpcBindAddr,
 		grpcGracePeriod,
 		grpcTLSSrvCert,
 		grpcTLSSrvKey,
-		grpcTLSSrvClientCA
+		grpcTLSSrvClientCA,
+		grpcMaxConnectionAge
 }
 
 func regHTTPFlags(cmd *kingpin.CmdClause) (httpBindAddr *string, httpGracePeriod *model.Duration) {

cmd/thanos/query.go

@@ -47,7 +47,7 @@ func registerQuery(m map[string]setupFunc, app *kingpin.Application) {
 	cmd := app.Command(comp.String(), "query node exposing PromQL enabled Query API with data retrieved from multiple store nodes")
 
 	httpBindAddr, httpGracePeriod := regHTTPFlags(cmd)
-	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA := regGRPCFlags(cmd)
+	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA, grpcMaxConnAge := regGRPCFlags(cmd)
 
 	secure := cmd.Flag("grpc-client-tls-secure", "Use TLS when talking to the gRPC server").Default("false").Bool()
 	cert := cmd.Flag("grpc-client-tls-cert", "TLS Certificates to use to identify this client to the server").Default("").String()
@@ -137,6 +137,7 @@ func registerQuery(m map[string]setupFunc, app *kingpin.Application) {
 			*grpcCert,
 			*grpcKey,
 			*grpcClientCA,
+			*grpcMaxConnAge,
 			*secure,
 			*cert,
 			*key,
@@ -219,6 +220,7 @@ func runQuery(
 	grpcCert string,
 	grpcKey string,
 	grpcClientCA string,
+	grpcMaxConnAge time.Duration,
 	secure bool,
 	cert string,
 	key string,
@@ -408,6 +410,7 @@ func runQuery(
 			grpcserver.WithListen(grpcBindAddr),
 			grpcserver.WithGracePeriod(grpcGracePeriod),
 			grpcserver.WithTLSConfig(tlsCfg),
+			grpcserver.WithMaxConnAge(grpcMaxConnAge),
 		)
 
 		g.Add(func() error {

cmd/thanos/receive.go

@@ -36,7 +36,7 @@ func registerReceive(m map[string]setupFunc, app *kingpin.Application) {
 	cmd := app.Command(comp.String(), "Accept Prometheus remote write API requests and write to local tsdb (EXPERIMENTAL, this may change drastically without notice)")
 
 	httpBindAddr, httpGracePeriod := regHTTPFlags(cmd)
-	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA := regGRPCFlags(cmd)
+	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA, grpcMaxConnAge := regGRPCFlags(cmd)
 
 	rwAddress := cmd.Flag("remote-write.address", "Address to listen on for remote write requests.").
 		Default("0.0.0.0:19291").String()
@@ -109,6 +109,7 @@ func registerReceive(m map[string]setupFunc, app *kingpin.Application) {
 			*grpcCert,
 			*grpcKey,
 			*grpcClientCA,
+			*grpcMaxConnAge,
 			*httpBindAddr,
 			time.Duration(*httpGracePeriod),
 			*rwAddress,
@@ -144,6 +145,7 @@ func runReceive(
 	grpcCert string,
 	grpcKey string,
 	grpcClientCA string,
+	grpcMaxConnAge time.Duration,
 	httpBindAddr string,
 	httpGracePeriod time.Duration,
 	rwAddress string,
@@ -370,6 +372,7 @@ func runReceive(
 					grpcserver.WithListen(grpcBindAddr),
 					grpcserver.WithGracePeriod(grpcGracePeriod),
 					grpcserver.WithTLSConfig(tlsCfg),
+					grpcserver.WithMaxConnAge(grpcMaxConnAge),
 				)
 				startGRPC <- struct{}{}
 			}

cmd/thanos/rule.go

@@ -63,7 +63,7 @@ func registerRule(m map[string]setupFunc, app *kingpin.Application) {
 	cmd := app.Command(comp.String(), "ruler evaluating Prometheus rules against given Query nodes, exposing Store API and storing old blocks in bucket")
 
 	httpBindAddr, httpGracePeriod := regHTTPFlags(cmd)
-	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA := regGRPCFlags(cmd)
+	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA, grpcMaxConnAge := regGRPCFlags(cmd)
 
 	labelStrs := cmd.Flag("label", "Labels to be applied to all generated metrics (repeated). Similar to external labels for Prometheus, used to identify ruler and its blocks as unique source.").
 		PlaceHolder("<name>=\"<value>\"").Strings()
@@ -162,6 +162,7 @@ func registerRule(m map[string]setupFunc, app *kingpin.Application) {
 			*grpcCert,
 			*grpcKey,
 			*grpcClientCA,
+			*grpcMaxConnAge,
 			*httpBindAddr,
 			time.Duration(*httpGracePeriod),
 			*webRoutePrefix,
@@ -199,6 +200,7 @@ func runRule(
 	grpcCert string,
 	grpcKey string,
 	grpcClientCA string,
+	grpcMaxConnAge time.Duration,
 	httpBindAddr string,
 	httpGracePeriod time.Duration,
 	webRoutePrefix string,
@@ -514,6 +516,7 @@ func runRule(
 			grpcserver.WithListen(grpcBindAddr),
 			grpcserver.WithGracePeriod(grpcGracePeriod),
 			grpcserver.WithTLSConfig(tlsCfg),
+			grpcserver.WithMaxConnAge(grpcMaxConnAge),
 		)
 
 		g.Add(func() error {

cmd/thanos/sidecar.go

@@ -37,7 +37,7 @@ func registerSidecar(m map[string]setupFunc, app *kingpin.Application) {
 	cmd := app.Command(component.Sidecar.String(), "sidecar for Prometheus server")
 
 	httpBindAddr, httpGracePeriod := regHTTPFlags(cmd)
-	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA := regGRPCFlags(cmd)
+	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA, grpcMaxConnAge := regGRPCFlags(cmd)
 
 	promURL := cmd.Flag("prometheus.url", "URL at which to reach Prometheus's API. For better performance use local network.").
 		Default("http://localhost:9090").URL()
@@ -82,6 +82,7 @@ func registerSidecar(m map[string]setupFunc, app *kingpin.Application) {
 			*grpcCert,
 			*grpcKey,
 			*grpcClientCA,
+			*grpcMaxConnAge,
 			*httpBindAddr,
 			time.Duration(*httpGracePeriod),
 			*promURL,
@@ -106,6 +107,7 @@ func runSidecar(
 	grpcCert string,
 	grpcKey string,
 	grpcClientCA string,
+	grpcMaxConnAge time.Duration,
 	httpBindAddr string,
 	httpGracePeriod time.Duration,
 	promURL *url.URL,
@@ -254,6 +256,7 @@ func runSidecar(
 			grpcserver.WithListen(grpcBindAddr),
 			grpcserver.WithGracePeriod(grpcGracePeriod),
 			grpcserver.WithTLSConfig(tlsCfg),
+			grpcserver.WithMaxConnAge(grpcMaxConnAge),
 		)
 		g.Add(func() error {
 			statusProber.Ready()

cmd/thanos/store.go

@@ -31,7 +31,7 @@ func registerStore(m map[string]setupFunc, app *kingpin.Application) {
 	cmd := app.Command(component.Store.String(), "store node giving access to blocks in a bucket provider. Now supported GCS, S3, Azure, Swift and Tencent COS.")
 
 	httpBindAddr, httpGracePeriod := regHTTPFlags(cmd)
-	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA := regGRPCFlags(cmd)
+	grpcBindAddr, grpcGracePeriod, grpcCert, grpcKey, grpcClientCA, grpcMaxConnAge := regGRPCFlags(cmd)
 
 	dataDir := cmd.Flag("data-dir", "Data directory in which to cache remote blocks.").
 		Default("./data").String()
@@ -84,6 +84,7 @@ func registerStore(m map[string]setupFunc, app *kingpin.Application) {
 			*grpcCert,
 			*grpcKey,
 			*grpcClientCA,
+			*grpcMaxConnAge,
 			*httpBindAddr,
 			time.Duration(*httpGracePeriod),
 			uint64(*indexCacheSize),
@@ -117,6 +118,7 @@ func runStore(
 	grpcCert string,
 	grpcKey string,
 	grpcClientCA string,
+	grpcMaxConnAge time.Duration,
 	httpBindAddr string,
 	httpGracePeriod time.Duration,
 	indexCacheSizeBytes uint64,
@@ -246,6 +248,7 @@ func runStore(
 			grpcserver.WithListen(grpcBindAddr),
 			grpcserver.WithGracePeriod(grpcGracePeriod),
 			grpcserver.WithTLSConfig(tlsCfg),
+			grpcserver.WithMaxConnAge(grpcMaxConnAge),
 		)
 
 		g.Add(func() error {

pkg/server/grpc/grpc.go

@@ -20,6 +20,7 @@ import (
 	"google.golang.org/grpc"
 	"google.golang.org/grpc/codes"
 	"google.golang.org/grpc/credentials"
+	"google.golang.org/grpc/keepalive"
 	"google.golang.org/grpc/status"
 )
 
@@ -74,6 +75,9 @@ func New(logger log.Logger, reg prometheus.Registerer, tracer opentracing.Tracer
 	if options.tlsConfig != nil {
 		grpcOpts = append(grpcOpts, grpc.Creds(credentials.NewTLS(options.tlsConfig)))
 	}
+	if options.maxConnAge > 0 {
+		grpcOpts = append(grpcOpts, grpc.KeepaliveParams(keepalive.ServerParameters{MaxConnectionAge: options.maxConnAge}))
+	}
 	s := grpc.NewServer(grpcOpts...)
 
 	storepb.RegisterStoreServer(s, storeSrv)

pkg/server/grpc/option.go

@@ -7,6 +7,7 @@ import (
 
 type options struct {
 	gracePeriod time.Duration
+	maxConnAge  time.Duration
 	listen      string
 
 	tlsConfig *tls.Config
@@ -45,3 +46,10 @@ func WithTLSConfig(cfg *tls.Config) Option {
 		o.tlsConfig = cfg
 	})
 }
+
+// WithMaxConnAge sets the maximum connection age for gRPC server.
+func WithMaxConnAge(t time.Duration) Option {
+	return optionFunc(func(o *options) {
+		o.maxConnAge = t
+	})
+}