[Promxy] Update promxy on recent prometheus and add native histogram support

Dockerfile

@@ -1,4 +1,4 @@
-FROM --platform=$BUILDPLATFORM golang:1.21-alpine3.18 as builder
+FROM --platform=$BUILDPLATFORM golang:1.22-alpine3.20 as builder
 
 ARG BUILDPLATFORM
 ARG TARGETARCH

Makefile

@@ -4,6 +4,9 @@ GOFILES := $(shell find . -name "*.go" -type f ! -path "./vendor/*")
 GOFMT ?= gofmt
 GOIMPORTS ?= goimports -local=github.com/jacksontj/promxy
 STATICCHECK ?= staticcheck
+GIT_BRANCH := $(shell git rev-parse --abbrev-ref HEAD)
+GIT_REVISION := $(shell git rev-parse --short HEAD)
+IMAGE_TAG ?= $(subst /,-,$(GIT_BRANCH))-$(GIT_REVISION)
 
 .PHONY: clean
 clean:
@@ -31,16 +34,20 @@ release:
 	./build.bash github.com/jacksontj/promxy/cmd/promxy $(BUILD)
 	./build.bash github.com/jacksontj/promxy/cmd/remote_write_exporter $(BUILD)
 
+.PHONY: build-image
+build-image: clean
+	docker buildx build --load --platform linux/amd64 --build-arg revision=$(GIT_REVISION) -t promxy:latest -t promxy:$(IMAGE_TAG) .
+
 testlocal-build:
 	docker build -t 127.0.0.1:32000/promxy:latest .
 	docker push 127.0.0.1:32000/promxy:latest
 
 .PHONY: vendor
 vendor:
-	GO111MODULE=on $(GO) mod tidy -compat=1.20
+	GO111MODULE=on $(GO) mod tidy -compat=1.22
 	GO111MODULE=on $(GO) mod vendor
 
 .PHONY: update-prom-fork
 update-prom-fork:
-	GO111MODULE=on $(GO) mod edit -replace github.com/prometheus/prometheus=github.com/jacksontj/prometheus@v0.2.37.5-fork
+	GO111MODULE=on $(GO) mod edit -replace github.com/prometheus/prometheus=github.com/rishabhkumar92/prometheus@vrishabhk-promxy-upgrade
 	$(MAKE) vendor

cmd/promxy/Makefile

@@ -6,4 +6,4 @@ test-local:
 	go build -race -mod=vendor -tags netgo,builtinassets && ./promxy --http.shutdown-delay=0s --config=local_test.conf --web.external-url=http://localhost:8082/promxy --log-level=debug
 
 version:
-	go build -race -mod=vendor -tags netgo,builtinassets && ./promxy --version
+	go build -race -mod=readonly -tags netgo,builtinassets && ./promxy --version

cmd/promxy/config.yaml

@@ -1,62 +1,64 @@
 ##
 ## Regular prometheus configuration
 ##
-global:
-  evaluation_interval: 5s
-  external_labels:
-    source: promxy
+
+# Uncomment below section to enable evaluation interval.
+#global:
+#  evaluation_interval: 5s
+#  external_labels:
+#    source: promxy
 
 # Rule files specifies a list of globs. Rules and alerts are read from
-# all matching files.
-rule_files:
-- "*rule"
+# all matching files. Uncomment below section to enable.
+
+#rule_files:
+#  - "*rule"
 
-# Alerting specifies settings related to the Alertmanager.
-alerting:
-  alertmanagers:
-  - scheme: http
-    static_configs:
-    - targets:
-      - "127.0.0.1:12345"
+# Alerting specifies settings related to the Alertmanager. Uncomment below section to enable.
+
+#alerting:
+#  alertmanagers:
+#    - scheme: http
+#      static_configs:
+#        - targets:
+#            - "127.0.0.1:12345"
 
 # remote_write configuration is used by promxy as its local Appender, meaning all
 # metrics promxy would "write" (not export) would be sent to this. Examples
-# of this include: recording rules, metrics on alerting rules, etc.
-remote_write:
-  - url: http://localhost:8083/receive
+# of this include: recording rules, metrics on alerting rules, etc. Uncomment below section to enable.
+
+#remote_write:
+#  - url: http://localhost:8083/receive
 
 ##
 ### Promxy configuration
 ##
 promxy:
+  # In below examples, point 8000 & 8001 to prometheus servers.
   server_groups:
     # All upstream prometheus service discovery mechanisms are supported with the same
     # markup, all defined in https://github.com/prometheus/prometheus/blob/master/discovery/config/config.go#L33
     - static_configs:
         - targets:
-          - localhost:9090
+            - localhost:8000
       # labels to be added to metrics retrieved from this server_group
       labels:
-        sg: localhost_9090
+        sg: localhost_8000
       # anti-affinity for merging values in timeseries between hosts in the server_group
       anti_affinity: 10s
       # time to wait for a server's response headers after fully writing the request (including its body, if any).
       # This time does not include the time to read the response body.
       timeout: 5s
       # Controls whether to use remote_read or the prom API for fetching remote RAW data (e.g. matrix selectors)
       # Note, some prometheus implementations (e.g. [VictoriaMetrics](https://github.com/prometheus/prometheus/issues/4456) don't support remote_read.
-      remote_read: true
+      remote_read: false
       # configures the path to send remote read requests to. The default is "api/v1/read"
-      remote_read_path: api/v1/read
       # path_prefix defines a prefix to prepend to all queries to hosts in this servergroup
       # This can be relabeled using __path_prefix__
-      path_prefix: /example/prefix
+      path_prefix: /prometheus
       # query_params adds the following map of query parameters to downstream requests.
       # The initial use-case for this is to add `nocache=1` to VictoriaMetrics downstreams
       # (see https://github.com/jacksontj/promxy/issues/202)
-      query_params:
-        nocache: 1
-      # configures the protocol scheme used for requests. Defaults to http
       scheme: http
       # options for promxy's HTTP client when talking to hosts in server_groups
       http_client:
@@ -65,40 +67,5 @@ promxy:
         dial_timeout: 1s
         tls_config:
           insecure_skip_verify: true
-
-      # relative_time_range defines a relative-to-now time range that this server group contains.
-      # this is completely optional and start/end are both optional as well
-      # an example is if this servergroup only has the most recent 3h of data
-      # the "start" would be -3h and the end would be left out
-      relative_time_range:
-        start: -3h
-        end: -1h
-        truncate: false
-
-      # when merging sample streams, the max value at a given timestamp will be preferred
-      prefer_max: false
-
-      # absolute_time_range defines an absolute time range that this server group contains.
-      # this is completely optional and start/end are both optional as well
-      # and example is if the servergroup has been deprecated and is no longer receiving data
-      # you could set the specific times that it has data for.
-      absolute_time_range:
-        start: '2009-10-10T23:00:00Z'
-        end: '2009-10-11T23:00:00Z'
-        truncate: true
-
-    # as many additional server groups as you have
-    - static_configs:
-        - targets:
-          - localhost:9091
-      labels:
-        sg: localhost_9091
-      anti_affinity: 10s
-      scheme: http
-      http_client:
-        tls_config:
-          insecure_skip_verify: true
-      # ignore_error will make the given server group's response "optional"
-      # meaning if this servergroup returns and error and others don't the overall
-      # query can still succeed
-      ignore_error: true
+      # make response as optional
+      ignore_error: false

cmd/promxy/main.go

@@ -7,6 +7,7 @@ import (
 	"io"
 	"net"
 	"net/http"
+	_ "net/http/pprof"
 	"net/url"
 	"os"
 	"os/signal"
@@ -16,18 +17,16 @@ import (
 	"syscall"
 	"time"
 
-	_ "net/http/pprof"
-
-	kitlog "github.com/go-kit/log"
 	"github.com/golang/glog"
 	"github.com/grafana/regexp"
 	"github.com/jessevdk/go-flags"
 	"github.com/julienschmidt/httprouter"
 	"github.com/prometheus/client_golang/prometheus"
+	collectorVersion "github.com/prometheus/client_golang/prometheus/collectors/version"
 	"github.com/prometheus/client_golang/prometheus/promauto"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
 	"github.com/prometheus/common/model"
-	"github.com/prometheus/common/promlog"
+	"github.com/prometheus/common/promslog"
 	"github.com/prometheus/common/version"
 	"github.com/prometheus/prometheus/config"
 	"github.com/prometheus/prometheus/discovery"
@@ -70,7 +69,7 @@ var (
 )
 
 func init() {
-	prometheus.MustRegister(version.NewCollector("promxy"))
+	prometheus.MustRegister(collectorVersion.NewCollector("promxy"))
 }
 
 type cliOpts struct {
@@ -109,7 +108,7 @@ type cliOpts struct {
 	ResendDelay               time.Duration `long:"rules.alert.resend-delay" description:"Minimum amount of time to wait before resending an alert to Alertmanager." default:"1m"`
 	AlertBackfill             bool          `long:"rules.alertbackfill" description:"Enable promxy to recalculate alert state on startup when the downstream datastore doesn't have an ALERTS_FOR_STATE"`
 
-	ShutdownDelay   time.Duration `long:"http.shutdown-delay" description:"time to wait before shutting down the http server, this allows for a grace period for upstreams (e.g. LoadBalancers) to discover the new stopping status through healthchecks" default:"10s"`
+	ShutdownDelay   time.Duration `long:"http.shutdown-delay" description:"time to wait before shutting down the http server, this allows for a grace period for upstreams (e.g. LoadBalancers) to discover the new stopping status through healthchecks" default:"1s"`
 	ShutdownTimeout time.Duration `long:"http.shutdown-timeout" description:"max time to wait for a graceful shutdown of the HTTP server" default:"60s"`
 }
 
@@ -238,15 +237,15 @@ func main() {
 	reloadables = append(reloadables, ps)
 	proxyStorage = ps
 
-	logCfg := &promlog.Config{
-		Level:  &promlog.AllowedLevel{},
-		Format: &promlog.AllowedFormat{},
+	logCfg := &promslog.Config{
+		Level:  &promslog.AllowedLevel{},
+		Format: &promslog.AllowedFormat{},
 	}
 	if err := logCfg.Level.Set("info"); err != nil {
 		logrus.Fatalf("Unable to set log level: %v", err)
 	}
 
-	logger := promlog.New(logCfg)
+	logger := promslog.New(logCfg)
 
 	engineOpts := promql.EngineOpts{
 		Reg:                      prometheus.DefaultRegisterer,
@@ -265,7 +264,7 @@ func main() {
 		if opts.LocalStoragePath == "" {
 			logrus.Fatalf("local storage path must be defined if you wish to enable max query concurrency limits")
 		}
-		engineOpts.ActiveQueryTracker = promql.NewActiveQueryTracker(opts.LocalStoragePath, opts.QueryMaxConcurrency, kitlog.With(logger, "component", "activeQueryTracker"))
+		engineOpts.ActiveQueryTracker = promql.NewActiveQueryTracker(opts.LocalStoragePath, opts.QueryMaxConcurrency, logger.With("component", "activeQueryTracker"))
 	}
 
 	engine := promql.NewEngine(engineOpts)
@@ -282,11 +281,11 @@ func main() {
 			Registerer:    prometheus.DefaultRegisterer,
 			QueueCapacity: opts.NotificationQueueCapacity,
 		},
-		kitlog.With(logger, "component", "notifier"),
+		logger.With("component", "notifier"),
 	)
 	reloadables = append(reloadables, proxyconfig.WrapPromReloadable(notifierManager))
 
-	discoveryManagerNotify := discovery.NewManager(ctx, kitlog.With(logger, "component", "discovery manager notify"))
+	discoveryManagerNotify := discovery.NewManager(ctx, logger.With("component", "discovery manager notify"), prometheus.DefaultRegisterer, nil, discovery.Name("notify"))
 
 	reloadables = append(reloadables,
 		proxyconfig.WrapPromReloadable(&proxyconfig.ApplyConfigFunc{func(cfg *config.Config) error {
@@ -387,7 +386,8 @@ func main() {
 	}}))
 
 	// We need an empty scrape manager, simply to make the API not panic and error out
-	scrapeManager := scrape.NewManager(nil, kitlog.With(logger, "component", "scrape manager"), nil)
+	scrapeManager, _ := scrape.NewManager(nil, logger.With("component", "scrape manager"), nil, proxyStorage, prometheus.DefaultRegisterer)
+	listenAddresses := []string{opts.BindAddr}
 
 	webOptions := &web.Options{
 		Registerer:      prometheus.DefaultRegisterer,
@@ -406,9 +406,10 @@ func main() {
 
 		EnableLifecycle: opts.EnableLifecycle,
 
-		Flags:       opts.ToFlags(),
-		RoutePrefix: opts.RoutePrefix,
-		ExternalURL: externalUrl,
+		Flags:           opts.ToFlags(),
+		ListenAddresses: listenAddresses,
+		RoutePrefix:     opts.RoutePrefix,
+		ExternalURL:     externalUrl,
 		Version: &web.PrometheusVersion{
 			Version:   version.Version,
 			Revision:  version.Revision,
@@ -433,7 +434,7 @@ func main() {
 
 	webHandler := web.New(logger, webOptions)
 	reloadables = append(reloadables, proxyconfig.WrapPromReloadable(webHandler))
-	webHandler.SetReady(true)
+	webHandler.SetReady(web.Ready)
 
 	apiPrefix := path.Join(webOptions.RoutePrefix, "/api/v1")
 	// Register API endpoint with correct route prefix