Upgrading server dependancies (#8308)

vendor/github.com/prometheus/client_golang/ISSUE_TEMPLATE.md

@@ -0,0 +1,8 @@
+<!--
+
+    If you are unhappy how your favorite Go dependency management tool deals
+    with this repository, please do not file an issue but read
+    https://github.com/prometheus/client_golang#important-note-about-releases-versioning-tagging-stability-and-your-favorite-go-dependency-management-tool
+    instead. Thank you very much.
+
+-->

vendor/github.com/prometheus/client_golang/README.md

@@ -2,6 +2,7 @@
 
 [![Build Status](https://travis-ci.org/prometheus/client_golang.svg?branch=master)](https://travis-ci.org/prometheus/client_golang)
 [![Go Report Card](https://goreportcard.com/badge/github.com/prometheus/client_golang)](https://goreportcard.com/report/github.com/prometheus/client_golang)
+[![go-doc](https://godoc.org/github.com/prometheus/client_golang?status.svg)](https://godoc.org/github.com/prometheus/client_golang)
 
 This is the [Go](http://golang.org) client library for
 [Prometheus](http://prometheus.io). It has two separate parts, one for
@@ -10,6 +11,50 @@ Prometheus HTTP API.
 
 __This library requires Go1.7 or later.__
 
+## Important note about releases, versioning, tagging, stability, and your favorite Go dependency management tool
+
+While our goal is to follow [Semantic Versioning](https://semver.org/), this
+repository is still pre-1.0.0. To quote the
+[Semantic Versioning spec](https://semver.org/#spec-item-4): “Anything may
+change at any time. The public API should not be considered stable.” We know
+that this is at odds with the widespread use of this library. However, just
+declaring something 1.0.0 doesn't make it 1.0.0. Instead, we are working
+towards a 1.0.0 release that actually deserves its major version number.
+
+Having said that, we aim for always keeping the tip of master in a workable
+state and for only introducing ”mildly” breaking changes up to and including
+[v0.9.0](https://github.com/prometheus/client_golang/milestone/1). After that,
+a number of ”hard” breaking changes are planned, see the
+[v0.10.0 milestone](https://github.com/prometheus/client_golang/milestone/2),
+which should get the library much closer to 1.0.0 state.
+
+Dependency management in Go projects is still in flux, and there are many tools
+floating around. While [dep](https://golang.github.io/dep/) might develop into
+the de-facto standard tool, it is still officially experimental. The roadmap
+for this library has been laid out with a lot of sometimes painful experience
+in mind. We really cannot adjust it every other month to the needs of the
+currently most popular or most promising Go dependency management tool. The
+recommended course of action with dependency management tools is the following:
+
+- Do not expect strict post-1.0.0 semver semantics prior to the 1.0.0
+  release. If your dependency management tool expects strict post-1.0.0 semver
+  semantics, you have to wait. Sorry.
+- If you want absolute certainty, please lock to a specific commit. You can
+  also lock to tags, but please don't ask for more tagging. This would suggest
+  some release or stability testing procedure that simply is not in place. As
+  said above, we are aiming for stability of the tip of master, but if we
+  tagged every single commit, locking to tags would be the same as locking to
+  commits.
+- If you want to get the newer features and improvements and are willing to
+  take the minor risk of newly introduced bugs and “mild” breakage, just always
+  update to the tip of master (which is essentially the original idea of Go
+  dependency management). We recommend to not use features marked as
+  _deprecated_ in this case.
+- Once [v0.9.0](https://github.com/prometheus/client_golang/milestone/1) is
+  out, you could lock to v0.9.x to get bugfixes (and perhaps minor new
+  features) while avoiding the “hard” breakage that will come with post-0.9
+  features.
+
 ## Instrumenting applications
 
 [![code-coverage](http://gocover.io/_badge/github.com/prometheus/client_golang/prometheus)](http://gocover.io/github.com/prometheus/client_golang/prometheus) [![go-doc](https://godoc.org/github.com/prometheus/client_golang/prometheus?status.svg)](https://godoc.org/github.com/prometheus/client_golang/prometheus)
@@ -26,7 +71,7 @@ contains simple examples of instrumented code.
 
 ## Client for the Prometheus HTTP API
 
-[![code-coverage](http://gocover.io/_badge/github.com/prometheus/client_golang/api/prometheus)](http://gocover.io/github.com/prometheus/client_golang/api/prometheus) [![go-doc](https://godoc.org/github.com/prometheus/client_golang/api/prometheus?status.svg)](https://godoc.org/github.com/prometheus/client_golang/api/prometheus)
+[![code-coverage](http://gocover.io/_badge/github.com/prometheus/client_golang/api/prometheus/v1)](http://gocover.io/github.com/prometheus/client_golang/api/prometheus/v1) [![go-doc](https://godoc.org/github.com/prometheus/client_golang/api/prometheus?status.svg)](https://godoc.org/github.com/prometheus/client_golang/api)
 
 The
 [`api/prometheus` directory](https://github.com/prometheus/client_golang/tree/master/api/prometheus)

vendor/github.com/prometheus/client_golang/prometheus/doc.go

@@ -11,10 +11,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-// Package prometheus provides metrics primitives to instrument code for
-// monitoring. It also offers a registry for metrics. Sub-packages allow to
-// expose the registered metrics via HTTP (package promhttp) or push them to a
-// Pushgateway (package push).
+// Package prometheus is the core instrumentation package. It provides metrics
+// primitives to instrument code for monitoring. It also offers a registry for
+// metrics. Sub-packages allow to expose the registered metrics via HTTP
+// (package promhttp) or push them to a Pushgateway (package push).
 //
 // All exported functions and methods are safe to be used concurrently unless
 // specified otherwise.

vendor/github.com/prometheus/client_golang/prometheus/http.go

@@ -61,9 +61,8 @@ func giveBuf(buf *bytes.Buffer) {
 // name).
 //
 // Deprecated: Please note the issues described in the doc comment of
-// InstrumentHandler. You might want to consider using promhttp.Handler instead
-// (which is not instrumented, but can be instrumented with the tooling provided
-// in package promhttp).
+// InstrumentHandler. You might want to consider using
+// promhttp.InstrumentedHandler instead.
 func Handler() http.Handler {
 	return InstrumentHandler("prometheus", UninstrumentedHandler())
 }

vendor/github.com/prometheus/client_golang/prometheus/process_collector.go

@@ -26,8 +26,11 @@ type processCollector struct {
 }
 
 // NewProcessCollector returns a collector which exports the current state of
-// process metrics including cpu, memory and file descriptor usage as well as
-// the process start time for the given process id under the given namespace.
+// process metrics including CPU, memory and file descriptor usage as well as
+// the process start time for the given process ID under the given namespace.
+//
+// Currently, the collector depends on a Linux-style proc filesystem and
+// therefore only exports metrics for Linux.
 func NewProcessCollector(pid int, namespace string) Collector {
 	return NewProcessCollectorPIDFn(
 		func() (int, error) { return pid, nil },
@@ -35,11 +38,8 @@ func NewProcessCollector(pid int, namespace string) Collector {
 	)
 }
 
-// NewProcessCollectorPIDFn returns a collector which exports the current state
-// of process metrics including cpu, memory and file descriptor usage as well
-// as the process start time under the given namespace. The given pidFn is
-// called on each collect and is used to determine the process to export
-// metrics for.
+// NewProcessCollectorPIDFn works like NewProcessCollector but the process ID is
+// determined on each collect anew by calling the given pidFn function.
 func NewProcessCollectorPIDFn(
 	pidFn func() (int, error),
 	namespace string,

vendor/github.com/prometheus/client_golang/prometheus/promhttp/delegator.go

@@ -102,10 +102,10 @@ func init() {
 		return d
 	}
 	pickDelegator[closeNotifier] = func(d *responseWriterDelegator) delegator { // 1
-		return closeNotifierDelegator{d}
+		return &closeNotifierDelegator{d}
 	}
 	pickDelegator[flusher] = func(d *responseWriterDelegator) delegator { // 2
-		return flusherDelegator{d}
+		return &flusherDelegator{d}
 	}
 	pickDelegator[flusher+closeNotifier] = func(d *responseWriterDelegator) delegator { // 3
 		return struct {
@@ -115,7 +115,7 @@ func init() {
 		}{d, &flusherDelegator{d}, &closeNotifierDelegator{d}}
 	}
 	pickDelegator[hijacker] = func(d *responseWriterDelegator) delegator { // 4
-		return hijackerDelegator{d}
+		return &hijackerDelegator{d}
 	}
 	pickDelegator[hijacker+closeNotifier] = func(d *responseWriterDelegator) delegator { // 5
 		return struct {

vendor/github.com/prometheus/client_golang/prometheus/promhttp/delegator_1_8.go

@@ -28,7 +28,7 @@ func (d *pusherDelegator) Push(target string, opts *http.PushOptions) error {
 
 func init() {
 	pickDelegator[pusher] = func(d *responseWriterDelegator) delegator { // 16
-		return pusherDelegator{d}
+		return &pusherDelegator{d}
 	}
 	pickDelegator[pusher+closeNotifier] = func(d *responseWriterDelegator) delegator { // 17
 		return struct {

vendor/github.com/prometheus/client_golang/prometheus/promhttp/http.go

@@ -39,6 +39,7 @@ import (
 	"net/http"
 	"strings"
 	"sync"
+	"time"
 
 	"github.com/prometheus/common/expfmt"
 
@@ -67,21 +68,51 @@ func giveBuf(buf *bytes.Buffer) {
 	bufPool.Put(buf)
 }
 
-// Handler returns an HTTP handler for the prometheus.DefaultGatherer. The
-// Handler uses the default HandlerOpts, i.e. report the first error as an HTTP
-// error, no error logging, and compression if requested by the client.
+// Handler returns an http.Handler for the prometheus.DefaultGatherer, using
+// default HandlerOpts, i.e. it reports the first error as an HTTP error, it has
+// no error logging, and it applies compression if requested by the client.
 //
-// If you want to create a Handler for the DefaultGatherer with different
-// HandlerOpts, create it with HandlerFor with prometheus.DefaultGatherer and
-// your desired HandlerOpts.
+// The returned http.Handler is already instrumented using the
+// InstrumentMetricHandler function and the prometheus.DefaultRegisterer. If you
+// create multiple http.Handlers by separate calls of the Handler function, the
+// metrics used for instrumentation will be shared between them, providing
+// global scrape counts.
+//
+// This function is meant to cover the bulk of basic use cases. If you are doing
+// anything that requires more customization (including using a non-default
+// Gatherer, different instrumentation, and non-default HandlerOpts), use the
+// HandlerFor function. See there for details.
 func Handler() http.Handler {
-	return HandlerFor(prometheus.DefaultGatherer, HandlerOpts{})
+	return InstrumentMetricHandler(
+		prometheus.DefaultRegisterer, HandlerFor(prometheus.DefaultGatherer, HandlerOpts{}),
+	)
 }
 
-// HandlerFor returns an http.Handler for the provided Gatherer. The behavior
-// of the Handler is defined by the provided HandlerOpts.
+// HandlerFor returns an uninstrumented http.Handler for the provided
+// Gatherer. The behavior of the Handler is defined by the provided
+// HandlerOpts. Thus, HandlerFor is useful to create http.Handlers for custom
+// Gatherers, with non-default HandlerOpts, and/or with custom (or no)
+// instrumentation. Use the InstrumentMetricHandler function to apply the same
+// kind of instrumentation as it is used by the Handler function.
 func HandlerFor(reg prometheus.Gatherer, opts HandlerOpts) http.Handler {
-	return http.HandlerFunc(func(w http.ResponseWriter, req *http.Request) {
+	var inFlightSem chan struct{}
+	if opts.MaxRequestsInFlight > 0 {
+		inFlightSem = make(chan struct{}, opts.MaxRequestsInFlight)
+	}
+
+	h := http.HandlerFunc(func(w http.ResponseWriter, req *http.Request) {
+		if inFlightSem != nil {
+			select {
+			case inFlightSem <- struct{}{}: // All good, carry on.
+				defer func() { <-inFlightSem }()
+			default:
+				http.Error(w, fmt.Sprintf(
+					"Limit of concurrent requests reached (%d), try again later.", opts.MaxRequestsInFlight,
+				), http.StatusServiceUnavailable)
+				return
+			}
+		}
+
 		mfs, err := reg.Gather()
 		if err != nil {
 			if opts.ErrorLog != nil {
@@ -137,9 +168,70 @@ func HandlerFor(reg prometheus.Gatherer, opts HandlerOpts) http.Handler {
 		if encoding != "" {
 			header.Set(contentEncodingHeader, encoding)
 		}
-		w.Write(buf.Bytes())
+		if _, err := w.Write(buf.Bytes()); err != nil && opts.ErrorLog != nil {
+			opts.ErrorLog.Println("error while sending encoded metrics:", err)
+		}
 		// TODO(beorn7): Consider streaming serving of metrics.
 	})
+
+	if opts.Timeout <= 0 {
+		return h
+	}
+	return http.TimeoutHandler(h, opts.Timeout, fmt.Sprintf(
+		"Exceeded configured timeout of %v.\n",
+		opts.Timeout,
+	))
+}
+
+// InstrumentMetricHandler is usually used with an http.Handler returned by the
+// HandlerFor function. It instruments the provided http.Handler with two
+// metrics: A counter vector "promhttp_metric_handler_requests_total" to count
+// scrapes partitioned by HTTP status code, and a gauge
+// "promhttp_metric_handler_requests_in_flight" to track the number of
+// simultaneous scrapes. This function idempotently registers collectors for
+// both metrics with the provided Registerer. It panics if the registration
+// fails. The provided metrics are useful to see how many scrapes hit the
+// monitored target (which could be from different Prometheus servers or other
+// scrapers), and how often they overlap (which would result in more than one
+// scrape in flight at the same time). Note that the scrapes-in-flight gauge
+// will contain the scrape by which it is exposed, while the scrape counter will
+// only get incremented after the scrape is complete (as only then the status
+// code is known). For tracking scrape durations, use the
+// "scrape_duration_seconds" gauge created by the Prometheus server upon each
+// scrape.
+func InstrumentMetricHandler(reg prometheus.Registerer, handler http.Handler) http.Handler {
+	cnt := prometheus.NewCounterVec(
+		prometheus.CounterOpts{
+			Name: "promhttp_metric_handler_requests_total",
+			Help: "Total number of scrapes by HTTP status code.",
+		},
+		[]string{"code"},
+	)
+	// Initialize the most likely HTTP status codes.
+	cnt.WithLabelValues("200")
+	cnt.WithLabelValues("500")
+	cnt.WithLabelValues("503")
+	if err := reg.Register(cnt); err != nil {
+		if are, ok := err.(prometheus.AlreadyRegisteredError); ok {
+			cnt = are.ExistingCollector.(*prometheus.CounterVec)
+		} else {
+			panic(err)
+		}
+	}
+
+	gge := prometheus.NewGauge(prometheus.GaugeOpts{
+		Name: "promhttp_metric_handler_requests_in_flight",
+		Help: "Current number of scrapes being served.",
+	})
+	if err := reg.Register(gge); err != nil {
+		if are, ok := err.(prometheus.AlreadyRegisteredError); ok {
+			gge = are.ExistingCollector.(prometheus.Gauge)
+		} else {
+			panic(err)
+		}
+	}
+
+	return InstrumentHandlerCounter(cnt, InstrumentHandlerInFlight(gge, handler))
 }
 
 // HandlerErrorHandling defines how a Handler serving metrics will handle
@@ -183,6 +275,21 @@ type HandlerOpts struct {
 	// If DisableCompression is true, the handler will never compress the
 	// response, even if requested by the client.
 	DisableCompression bool
+	// The number of concurrent HTTP requests is limited to
+	// MaxRequestsInFlight. Additional requests are responded to with 503
+	// Service Unavailable and a suitable message in the body. If
+	// MaxRequestsInFlight is 0 or negative, no limit is applied.
+	MaxRequestsInFlight int
+	// If handling a request takes longer than Timeout, it is responded to
+	// with 503 ServiceUnavailable and a suitable Message. No timeout is
+	// applied if Timeout is 0 or negative. Note that with the current
+	// implementation, reaching the timeout simply ends the HTTP requests as
+	// described above (and even that only if sending of the body hasn't
+	// started yet), while the bulk work of gathering all the metrics keeps
+	// running in the background (with the eventual result to be thrown
+	// away). Until the implementation is improved, it is recommended to
+	// implement a separate timeout in potentially slow Collectors.
+	Timeout time.Duration
 }
 
 // decorateWriter wraps a writer to handle gzip compression if requested.  It

vendor/github.com/prometheus/client_golang/prometheus/promhttp/http_test.go

@@ -19,7 +19,9 @@ import (
 	"log"
 	"net/http"
 	"net/http/httptest"
+	"strings"
 	"testing"
+	"time"
 
 	"github.com/prometheus/client_golang/prometheus"
 )
@@ -37,6 +39,23 @@ func (e errorCollector) Collect(ch chan<- prometheus.Metric) {
 	)
 }
 
+type blockingCollector struct {
+	CollectStarted, Block chan struct{}
+}
+
+func (b blockingCollector) Describe(ch chan<- *prometheus.Desc) {
+	ch <- prometheus.NewDesc("dummy_desc", "not helpful", nil, nil)
+}
+
+func (b blockingCollector) Collect(ch chan<- prometheus.Metric) {
+	select {
+	case b.CollectStarted <- struct{}{}:
+	default:
+	}
+	// Collects nothing, just waits for a channel receive.
+	<-b.Block
+}
+
 func TestHandlerErrorHandling(t *testing.T) {
 
 	// Create a registry that collects a MetricFamily with two elements,
@@ -129,3 +148,103 @@ the_count 0
 	}()
 	panicHandler.ServeHTTP(writer, request)
 }
+
+func TestInstrumentMetricHandler(t *testing.T) {
+	reg := prometheus.NewRegistry()
+	handler := InstrumentMetricHandler(reg, HandlerFor(reg, HandlerOpts{}))
+	// Do it again to test idempotency.
+	InstrumentMetricHandler(reg, HandlerFor(reg, HandlerOpts{}))
+	writer := httptest.NewRecorder()
+	request, _ := http.NewRequest("GET", "/", nil)
+	request.Header.Add("Accept", "test/plain")
+
+	handler.ServeHTTP(writer, request)
+	if got, want := writer.Code, http.StatusOK; got != want {
+		t.Errorf("got HTTP status code %d, want %d", got, want)
+	}
+
+	want := "promhttp_metric_handler_requests_in_flight 1\n"
+	if got := writer.Body.String(); !strings.Contains(got, want) {
+		t.Errorf("got body %q, does not contain %q", got, want)
+	}
+	want = "promhttp_metric_handler_requests_total{code=\"200\"} 0\n"
+	if got := writer.Body.String(); !strings.Contains(got, want) {
+		t.Errorf("got body %q, does not contain %q", got, want)
+	}
+
+	writer.Body.Reset()
+	handler.ServeHTTP(writer, request)
+	if got, want := writer.Code, http.StatusOK; got != want {
+		t.Errorf("got HTTP status code %d, want %d", got, want)
+	}
+
+	want = "promhttp_metric_handler_requests_in_flight 1\n"
+	if got := writer.Body.String(); !strings.Contains(got, want) {
+		t.Errorf("got body %q, does not contain %q", got, want)
+	}
+	want = "promhttp_metric_handler_requests_total{code=\"200\"} 1\n"
+	if got := writer.Body.String(); !strings.Contains(got, want) {
+		t.Errorf("got body %q, does not contain %q", got, want)
+	}
+}
+
+func TestHandlerMaxRequestsInFlight(t *testing.T) {
+	reg := prometheus.NewRegistry()
+	handler := HandlerFor(reg, HandlerOpts{MaxRequestsInFlight: 1})
+	w1 := httptest.NewRecorder()
+	w2 := httptest.NewRecorder()
+	w3 := httptest.NewRecorder()
+	request, _ := http.NewRequest("GET", "/", nil)
+	request.Header.Add("Accept", "test/plain")
+
+	c := blockingCollector{Block: make(chan struct{}), CollectStarted: make(chan struct{}, 1)}
+	reg.MustRegister(c)
+
+	rq1Done := make(chan struct{})
+	go func() {
+		handler.ServeHTTP(w1, request)
+		close(rq1Done)
+	}()
+	<-c.CollectStarted
+
+	handler.ServeHTTP(w2, request)
+
+	if got, want := w2.Code, http.StatusServiceUnavailable; got != want {
+		t.Errorf("got HTTP status code %d, want %d", got, want)
+	}
+	if got, want := w2.Body.String(), "Limit of concurrent requests reached (1), try again later.\n"; got != want {
+		t.Errorf("got body %q, want %q", got, want)
+	}
+
+	close(c.Block)
+	<-rq1Done
+
+	handler.ServeHTTP(w3, request)
+
+	if got, want := w3.Code, http.StatusOK; got != want {
+		t.Errorf("got HTTP status code %d, want %d", got, want)
+	}
+}
+
+func TestHandlerTimeout(t *testing.T) {
+	reg := prometheus.NewRegistry()
+	handler := HandlerFor(reg, HandlerOpts{Timeout: time.Millisecond})
+	w := httptest.NewRecorder()
+
+	request, _ := http.NewRequest("GET", "/", nil)
+	request.Header.Add("Accept", "test/plain")
+
+	c := blockingCollector{Block: make(chan struct{}), CollectStarted: make(chan struct{}, 1)}
+	reg.MustRegister(c)
+
+	handler.ServeHTTP(w, request)
+
+	if got, want := w.Code, http.StatusServiceUnavailable; got != want {
+		t.Errorf("got HTTP status code %d, want %d", got, want)
+	}
+	if got, want := w.Body.String(), "Exceeded configured timeout of 1ms.\n"; got != want {
+		t.Errorf("got body %q, want %q", got, want)
+	}
+
+	close(c.Block) // To not leak a goroutine.
+}

vendor/github.com/prometheus/client_golang/prometheus/promhttp/instrument_server_test.go

@@ -281,6 +281,16 @@ func (t *testResponseWriter) ReadFrom(io.Reader) (int64, error) {
 	return 0, nil
 }
 
+// testFlusher is an http.ResponseWriter that also implements http.Flusher.
+type testFlusher struct {
+	flushCalled bool
+}
+
+func (t *testFlusher) Header() http.Header       { return nil }
+func (t *testFlusher) Write([]byte) (int, error) { return 0, nil }
+func (t *testFlusher) WriteHeader(int)           {}
+func (t *testFlusher) Flush()                    { t.flushCalled = true }
+
 func TestInterfaceUpgrade(t *testing.T) {
 	w := &testResponseWriter{}
 	d := newDelegator(w, nil)
@@ -299,6 +309,22 @@ func TestInterfaceUpgrade(t *testing.T) {
 	if _, ok := d.(http.Hijacker); ok {
 		t.Error("delegator unexpectedly implements http.Hijacker")
 	}
+
+	f := &testFlusher{}
+	d = newDelegator(f, nil)
+	if _, ok := d.(http.CloseNotifier); ok {
+		t.Error("delegator unexpectedly implements http.CloseNotifier")
+	}
+	d.(http.Flusher).Flush()
+	if !w.flushCalled {
+		t.Error("Flush not called")
+	}
+	if _, ok := d.(io.ReaderFrom); ok {
+		t.Error("delegator unexpectedly implements io.ReaderFrom")
+	}
+	if _, ok := d.(http.Hijacker); ok {
+		t.Error("delegator unexpectedly implements http.Hijacker")
+	}
 }
 
 func ExampleInstrumentHandlerDuration() {

vendor/github.com/prometheus/client_golang/prometheus/push/deprecated.go

@@ -0,0 +1,172 @@
+// Copyright 2018 The Prometheus Authors
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package push
+
+// This file contains only deprecated code. Remove after v0.9 is released.
+
+import (
+	"bytes"
+	"fmt"
+	"io/ioutil"
+	"net/http"
+	"net/url"
+	"os"
+	"strings"
+
+	"github.com/prometheus/common/expfmt"
+	"github.com/prometheus/common/model"
+
+	"github.com/prometheus/client_golang/prometheus"
+)
+
+// FromGatherer triggers a metric collection by the provided Gatherer (which is
+// usually implemented by a prometheus.Registry) and pushes all gathered metrics
+// to the Pushgateway specified by url, using the provided job name and the
+// (optional) further grouping labels (the grouping map may be nil). See the
+// Pushgateway documentation for detailed implications of the job and other
+// grouping labels. Neither the job name nor any grouping label value may
+// contain a "/". The metrics pushed must not contain a job label of their own
+// nor any of the grouping labels.
+//
+// You can use just host:port or ip:port as url, in which case 'http://' is
+// added automatically. You can also include the schema in the URL. However, do
+// not include the '/metrics/jobs/...' part.
+//
+// Note that all previously pushed metrics with the same job and other grouping
+// labels will be replaced with the metrics pushed by this call. (It uses HTTP
+// method 'PUT' to push to the Pushgateway.)
+//
+// Deprecated: Please use a Pusher created with New instead.
+func FromGatherer(job string, grouping map[string]string, url string, g prometheus.Gatherer) error {
+	return push(job, grouping, url, g, "PUT")
+}
+
+// AddFromGatherer works like FromGatherer, but only previously pushed metrics
+// with the same name (and the same job and other grouping labels) will be
+// replaced. (It uses HTTP method 'POST' to push to the Pushgateway.)
+//
+// Deprecated: Please use a Pusher created with New instead.
+func AddFromGatherer(job string, grouping map[string]string, url string, g prometheus.Gatherer) error {
+	return push(job, grouping, url, g, "POST")
+}
+
+func push(job string, grouping map[string]string, pushURL string, g prometheus.Gatherer, method string) error {
+	if !strings.Contains(pushURL, "://") {
+		pushURL = "http://" + pushURL
+	}
+	if strings.HasSuffix(pushURL, "/") {
+		pushURL = pushURL[:len(pushURL)-1]
+	}
+
+	if strings.Contains(job, "/") {
+		return fmt.Errorf("job contains '/': %s", job)
+	}
+	urlComponents := []string{url.QueryEscape(job)}
+	for ln, lv := range grouping {
+		if !model.LabelName(ln).IsValid() {
+			return fmt.Errorf("grouping label has invalid name: %s", ln)
+		}
+		if strings.Contains(lv, "/") {
+			return fmt.Errorf("value of grouping label %s contains '/': %s", ln, lv)
+		}
+		urlComponents = append(urlComponents, ln, lv)
+	}
+	pushURL = fmt.Sprintf("%s/metrics/job/%s", pushURL, strings.Join(urlComponents, "/"))
+
+	mfs, err := g.Gather()
+	if err != nil {
+		return err
+	}
+	buf := &bytes.Buffer{}
+	enc := expfmt.NewEncoder(buf, expfmt.FmtProtoDelim)
+	// Check for pre-existing grouping labels:
+	for _, mf := range mfs {
+		for _, m := range mf.GetMetric() {
+			for _, l := range m.GetLabel() {
+				if l.GetName() == "job" {
+					return fmt.Errorf("pushed metric %s (%s) already contains a job label", mf.GetName(), m)
+				}
+				if _, ok := grouping[l.GetName()]; ok {
+					return fmt.Errorf(
+						"pushed metric %s (%s) already contains grouping label %s",
+						mf.GetName(), m, l.GetName(),
+					)
+				}
+			}
+		}
+		enc.Encode(mf)
+	}
+	req, err := http.NewRequest(method, pushURL, buf)
+	if err != nil {
+		return err
+	}
+	req.Header.Set(contentTypeHeader, string(expfmt.FmtProtoDelim))
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return err
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != 202 {
+		body, _ := ioutil.ReadAll(resp.Body) // Ignore any further error as this is for an error message only.
+		return fmt.Errorf("unexpected status code %d while pushing to %s: %s", resp.StatusCode, pushURL, body)
+	}
+	return nil
+}
+
+// Collectors works like FromGatherer, but it does not use a Gatherer. Instead,
+// it collects from the provided collectors directly. It is a convenient way to
+// push only a few metrics.
+//
+// Deprecated: Please use a Pusher created with New instead.
+func Collectors(job string, grouping map[string]string, url string, collectors ...prometheus.Collector) error {
+	return pushCollectors(job, grouping, url, "PUT", collectors...)
+}
+
+// AddCollectors works like AddFromGatherer, but it does not use a Gatherer.
+// Instead, it collects from the provided collectors directly. It is a
+// convenient way to push only a few metrics.
+//
+// Deprecated: Please use a Pusher created with New instead.
+func AddCollectors(job string, grouping map[string]string, url string, collectors ...prometheus.Collector) error {
+	return pushCollectors(job, grouping, url, "POST", collectors...)
+}
+
+func pushCollectors(job string, grouping map[string]string, url, method string, collectors ...prometheus.Collector) error {
+	r := prometheus.NewRegistry()
+	for _, collector := range collectors {
+		if err := r.Register(collector); err != nil {
+			return err
+		}
+	}
+	return push(job, grouping, url, r, method)
+}
+
+// HostnameGroupingKey returns a label map with the only entry
+// {instance="<hostname>"}. This can be conveniently used as the grouping
+// parameter if metrics should be pushed with the hostname as label. The
+// returned map is created upon each call so that the caller is free to add more
+// labels to the map.
+//
+// Deprecated: Usually, metrics pushed to the Pushgateway should not be
+// host-centric. (You would use https://github.com/prometheus/node_exporter in
+// that case.) If you have the need to add the hostname to the grouping key, you
+// are probably doing something wrong. See
+// https://prometheus.io/docs/practices/pushing/ for details.
+func HostnameGroupingKey() map[string]string {
+	hostname, err := os.Hostname()
+	if err != nil {
+		return map[string]string{"instance": "unknown"}
+	}
+	return map[string]string{"instance": hostname}
+}

vendor/github.com/prometheus/client_golang/prometheus/push/example_add_from_gatherer_test.go

@@ -11,12 +11,6 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-// Copyright (c) 2013, The Prometheus Authors
-// All rights reserved.
-//
-// Use of this source code is governed by a BSD-style license that can be found
-// in the LICENSE file.
-
 package push_test
 
 import (
@@ -53,10 +47,14 @@ func performBackup() (int, error) {
 	return 42, nil
 }
 
-func ExampleAddFromGatherer() {
+func ExamplePusher_Add() {
+	// We use a registry here to benefit from the consistency checks that
+	// happen during registration.
 	registry := prometheus.NewRegistry()
 	registry.MustRegister(completionTime, duration, records)
-	// Note that successTime is not registered at this time.
+	// Note that successTime is not registered.
+
+	pusher := push.New("http://pushgateway:9091", "db_backup").Gatherer(registry)
 
 	start := time.Now()
 	n, err := performBackup()
@@ -67,18 +65,16 @@ func ExampleAddFromGatherer() {
 	if err != nil {
 		fmt.Println("DB backup failed:", err)
 	} else {
-		// Only now register successTime.
-		registry.MustRegister(successTime)
+		// Add successTime to pusher only in case of success.
+		// We could as well register it with the registry.
+		// This example, however, demonstrates that you can
+		// mix Gatherers and Collectors when handling a Pusher.
+		pusher.Collector(successTime)
 		successTime.SetToCurrentTime()
 	}
-	// AddFromGatherer is used here rather than FromGatherer to not delete a
-	// previously pushed success timestamp in case of a failure of this
-	// backup.
-	if err := push.AddFromGatherer(
-		"db_backup", nil,
-		"http://pushgateway:9091",
-		registry,
-	); err != nil {
+	// Add is used here rather than Push to not delete a previously pushed
+	// success timestamp in case of a failure of this backup.
+	if err := pusher.Add(); err != nil {
 		fmt.Println("Could not push to Pushgateway:", err)
 	}
 }

vendor/github.com/prometheus/client_golang/prometheus/push/examples_test.go

@@ -20,17 +20,16 @@ import (
 	"github.com/prometheus/client_golang/prometheus/push"
 )
 
-func ExampleCollectors() {
+func ExamplePusher_Push() {
 	completionTime := prometheus.NewGauge(prometheus.GaugeOpts{
 		Name: "db_backup_last_completion_timestamp_seconds",
 		Help: "The timestamp of the last successful completion of a DB backup.",
 	})
 	completionTime.SetToCurrentTime()
-	if err := push.Collectors(
-		"db_backup", push.HostnameGroupingKey(),
-		"http://pushgateway:9091",
-		completionTime,
-	); err != nil {
+	if err := push.New("http://pushgateway:9091", "db_backup").
+		Collector(completionTime).
+		Grouping("db", "customers").
+		Push(); err != nil {
 		fmt.Println("Could not push completion time to Pushgateway:", err)
 	}
 }

vendor/github.com/prometheus/client_golang/prometheus/push/push.go

@@ -11,20 +11,27 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-// Copyright (c) 2013, The Prometheus Authors
-// All rights reserved.
+// Package push provides functions to push metrics to a Pushgateway. It uses a
+// builder approach. Create a Pusher with New and then add the various options
+// by using its methods, finally calling Add or Push, like this:
 //
-// Use of this source code is governed by a BSD-style license that can be found
-// in the LICENSE file.
-
-// Package push provides functions to push metrics to a Pushgateway. The metrics
-// to push are either collected from a provided registry, or from explicitly
-// listed collectors.
+//    // Easy case:
+//    push.New("http://example.org/metrics", "my_job").Gatherer(myRegistry).Push()
 //
-// See the documentation of the Pushgateway to understand the meaning of the
-// grouping parameters and the differences between push.Registry and
-// push.Collectors on the one hand and push.AddRegistry and push.AddCollectors
-// on the other hand: https://github.com/prometheus/pushgateway
+//    // Complex case:
+//    push.New("http://example.org/metrics", "my_job").
+//        Collector(myCollector1).
+//        Collector(myCollector2).
+//        Grouping("zone", "xy").
+//        Client(&myHTTPClient).
+//        BasicAuth("top", "secret").
+//        Add()
+//
+// See the examples section for more detailed examples.
+//
+// See the documentation of the Pushgateway to understand the meaning of
+// the grouping key and the differences between Push and Add:
+// https://github.com/prometheus/pushgateway
 package push
 
 import (
@@ -33,7 +40,6 @@ import (
 	"io/ioutil"
 	"net/http"
 	"net/url"
-	"os"
 	"strings"
 
 	"github.com/prometheus/common/expfmt"
@@ -44,57 +50,149 @@ import (
 
 const contentTypeHeader = "Content-Type"
 
-// FromGatherer triggers a metric collection by the provided Gatherer (which is
-// usually implemented by a prometheus.Registry) and pushes all gathered metrics
-// to the Pushgateway specified by url, using the provided job name and the
-// (optional) further grouping labels (the grouping map may be nil). See the
-// Pushgateway documentation for detailed implications of the job and other
-// grouping labels. Neither the job name nor any grouping label value may
-// contain a "/". The metrics pushed must not contain a job label of their own
-// nor any of the grouping labels.
-//
-// You can use just host:port or ip:port as url, in which case 'http://' is
-// added automatically. You can also include the schema in the URL. However, do
-// not include the '/metrics/jobs/...' part.
-//
-// Note that all previously pushed metrics with the same job and other grouping
-// labels will be replaced with the metrics pushed by this call. (It uses HTTP
-// method 'PUT' to push to the Pushgateway.)
-func FromGatherer(job string, grouping map[string]string, url string, g prometheus.Gatherer) error {
-	return push(job, grouping, url, g, "PUT")
+// Pusher manages a push to the Pushgateway. Use New to create one, configure it
+// with its methods, and finally use the Add or Push method to push.
+type Pusher struct {
+	error error
+
+	url, job string
+	grouping map[string]string
+
+	gatherers  prometheus.Gatherers
+	registerer prometheus.Registerer
+
+	client             *http.Client
+	useBasicAuth       bool
+	username, password string
 }
 
-// AddFromGatherer works like FromGatherer, but only previously pushed metrics
-// with the same name (and the same job and other grouping labels) will be
-// replaced. (It uses HTTP method 'POST' to push to the Pushgateway.)
-func AddFromGatherer(job string, grouping map[string]string, url string, g prometheus.Gatherer) error {
-	return push(job, grouping, url, g, "POST")
-}
-
-func push(job string, grouping map[string]string, pushURL string, g prometheus.Gatherer, method string) error {
-	if !strings.Contains(pushURL, "://") {
-		pushURL = "http://" + pushURL
+// New creates a new Pusher to push to the provided URL withe the provided job
+// name. You can use just host:port or ip:port as url, in which case “http://”
+// is added automatically. Alternatively, include the schema in the
+// URL. However, do not include the “/metrics/jobs/…” part.
+//
+// Note that until https://github.com/prometheus/pushgateway/issues/97 is
+// resolved, a “/” character in the job name is prohibited.
+func New(url, job string) *Pusher {
+	var (
+		reg = prometheus.NewRegistry()
+		err error
+	)
+	if !strings.Contains(url, "://") {
+		url = "http://" + url
 	}
-	if strings.HasSuffix(pushURL, "/") {
-		pushURL = pushURL[:len(pushURL)-1]
+	if strings.HasSuffix(url, "/") {
+		url = url[:len(url)-1]
 	}
-
 	if strings.Contains(job, "/") {
-		return fmt.Errorf("job contains '/': %s", job)
+		err = fmt.Errorf("job contains '/': %s", job)
 	}
-	urlComponents := []string{url.QueryEscape(job)}
-	for ln, lv := range grouping {
-		if !model.LabelName(ln).IsValid() {
-			return fmt.Errorf("grouping label has invalid name: %s", ln)
+
+	return &Pusher{
+		error:      err,
+		url:        url,
+		job:        job,
+		grouping:   map[string]string{},
+		gatherers:  prometheus.Gatherers{reg},
+		registerer: reg,
+		client:     &http.Client{},
+	}
+}
+
+// Push collects/gathers all metrics from all Collectors and Gatherers added to
+// this Pusher. Then, it pushes them to the Pushgateway configured while
+// creating this Pusher, using the configured job name and any added grouping
+// labels as grouping key. All previously pushed metrics with the same job and
+// other grouping labels will be replaced with the metrics pushed by this
+// call. (It uses HTTP method “PUT” to push to the Pushgateway.)
+//
+// Push returns the first error encountered by any method call (including this
+// one) in the lifetime of the Pusher.
+func (p *Pusher) Push() error {
+	return p.push("PUT")
+}
+
+// Add works like push, but only previously pushed metrics with the same name
+// (and the same job and other grouping labels) will be replaced. (It uses HTTP
+// method “POST” to push to the Pushgateway.)
+func (p *Pusher) Add() error {
+	return p.push("POST")
+}
+
+// Gatherer adds a Gatherer to the Pusher, from which metrics will be gathered
+// to push them to the Pushgateway. The gathered metrics must not contain a job
+// label of their own.
+//
+// For convenience, this method returns a pointer to the Pusher itself.
+func (p *Pusher) Gatherer(g prometheus.Gatherer) *Pusher {
+	p.gatherers = append(p.gatherers, g)
+	return p
+}
+
+// Collector adds a Collector to the Pusher, from which metrics will be
+// collected to push them to the Pushgateway. The collected metrics must not
+// contain a job label of their own.
+//
+// For convenience, this method returns a pointer to the Pusher itself.
+func (p *Pusher) Collector(c prometheus.Collector) *Pusher {
+	if p.error == nil {
+		p.error = p.registerer.Register(c)
+	}
+	return p
+}
+
+// Grouping adds a label pair to the grouping key of the Pusher, replacing any
+// previously added label pair with the same label name. Note that setting any
+// labels in the grouping key that are already contained in the metrics to push
+// will lead to an error.
+//
+// For convenience, this method returns a pointer to the Pusher itself.
+//
+// Note that until https://github.com/prometheus/pushgateway/issues/97 is
+// resolved, this method does not allow a “/” character in the label value.
+func (p *Pusher) Grouping(name, value string) *Pusher {
+	if p.error == nil {
+		if !model.LabelName(name).IsValid() {
+			p.error = fmt.Errorf("grouping label has invalid name: %s", name)
+			return p
 		}
-		if strings.Contains(lv, "/") {
-			return fmt.Errorf("value of grouping label %s contains '/': %s", ln, lv)
+		if strings.Contains(value, "/") {
+			p.error = fmt.Errorf("value of grouping label %s contains '/': %s", name, value)
+			return p
 		}
+		p.grouping[name] = value
+	}
+	return p
+}
+
+// Client sets a custom HTTP client for the Pusher. For convenience, this method
+// returns a pointer to the Pusher itself.
+func (p *Pusher) Client(c *http.Client) *Pusher {
+	p.client = c
+	return p
+}
+
+// BasicAuth configures the Pusher to use HTTP Basic Authentication with the
+// provided username and password. For convenience, this method returns a
+// pointer to the Pusher itself.
+func (p *Pusher) BasicAuth(username, password string) *Pusher {
+	p.useBasicAuth = true
+	p.username = username
+	p.password = password
+	return p
+}
+
+func (p *Pusher) push(method string) error {
+	if p.error != nil {
+		return p.error
+	}
+	urlComponents := []string{url.QueryEscape(p.job)}
+	for ln, lv := range p.grouping {
 		urlComponents = append(urlComponents, ln, lv)
 	}
-	pushURL = fmt.Sprintf("%s/metrics/job/%s", pushURL, strings.Join(urlComponents, "/"))
+	pushURL := fmt.Sprintf("%s/metrics/job/%s", p.url, strings.Join(urlComponents, "/"))
 
-	mfs, err := g.Gather()
+	mfs, err := p.gatherers.Gather()
 	if err != nil {
 		return err
 	}
@@ -107,7 +205,7 @@ func push(job string, grouping map[string]string, pushURL string, g prometheus.G
 				if l.GetName() == "job" {
 					return fmt.Errorf("pushed metric %s (%s) already contains a job label", mf.GetName(), m)
 				}
-				if _, ok := grouping[l.GetName()]; ok {
+				if _, ok := p.grouping[l.GetName()]; ok {
 					return fmt.Errorf(
 						"pushed metric %s (%s) already contains grouping label %s",
 						mf.GetName(), m, l.GetName(),
@@ -121,8 +219,11 @@ func push(job string, grouping map[string]string, pushURL string, g prometheus.G
 	if err != nil {
 		return err
 	}
+	if p.useBasicAuth {
+		req.SetBasicAuth(p.username, p.password)
+	}
 	req.Header.Set(contentTypeHeader, string(expfmt.FmtProtoDelim))
-	resp, err := http.DefaultClient.Do(req)
+	resp, err := p.client.Do(req)
 	if err != nil {
 		return err
 	}
@@ -133,40 +234,3 @@ func push(job string, grouping map[string]string, pushURL string, g prometheus.G
 	}
 	return nil
 }
-
-// Collectors works like FromGatherer, but it does not use a Gatherer. Instead,
-// it collects from the provided collectors directly. It is a convenient way to
-// push only a few metrics.
-func Collectors(job string, grouping map[string]string, url string, collectors ...prometheus.Collector) error {
-	return pushCollectors(job, grouping, url, "PUT", collectors...)
-}
-
-// AddCollectors works like AddFromGatherer, but it does not use a Gatherer.
-// Instead, it collects from the provided collectors directly. It is a
-// convenient way to push only a few metrics.
-func AddCollectors(job string, grouping map[string]string, url string, collectors ...prometheus.Collector) error {
-	return pushCollectors(job, grouping, url, "POST", collectors...)
-}
-
-func pushCollectors(job string, grouping map[string]string, url, method string, collectors ...prometheus.Collector) error {
-	r := prometheus.NewRegistry()
-	for _, collector := range collectors {
-		if err := r.Register(collector); err != nil {
-			return err
-		}
-	}
-	return push(job, grouping, url, r, method)
-}
-
-// HostnameGroupingKey returns a label map with the only entry
-// {instance="<hostname>"}. This can be conveniently used as the grouping
-// parameter if metrics should be pushed with the hostname as label. The
-// returned map is created upon each call so that the caller is free to add more
-// labels to the map.
-func HostnameGroupingKey() map[string]string {
-	hostname, err := os.Hostname()
-	if err != nil {
-		return map[string]string{"instance": "unknown"}
-	}
-	return map[string]string{"instance": hostname}
-}

vendor/github.com/prometheus/client_golang/prometheus/push/push_test.go

@@ -11,12 +11,6 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-// Copyright (c) 2013, The Prometheus Authors
-// All rights reserved.
-//
-// Use of this source code is governed by a BSD-style license that can be found
-// in the LICENSE file.
-
 package push
 
 import (
@@ -24,7 +18,6 @@ import (
 	"io/ioutil"
 	"net/http"
 	"net/http/httptest"
-	"os"
 	"testing"
 
 	"github.com/prometheus/common/expfmt"
@@ -40,11 +33,6 @@ func TestPush(t *testing.T) {
 		lastPath   string
 	)
 
-	host, err := os.Hostname()
-	if err != nil {
-		t.Error(err)
-	}
-
 	// Fake a Pushgateway that always responds with 202.
 	pgwOK := httptest.NewServer(
 		http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -98,26 +86,15 @@ func TestPush(t *testing.T) {
 	}
 	wantBody := buf.Bytes()
 
-	// PushCollectors, all good.
-	if err := Collectors("testjob", HostnameGroupingKey(), pgwOK.URL, metric1, metric2); err != nil {
+	// Push some Collectors, all good.
+	if err := New(pgwOK.URL, "testjob").
+		Collector(metric1).
+		Collector(metric2).
+		Push(); err != nil {
 		t.Fatal(err)
 	}
 	if lastMethod != "PUT" {
-		t.Error("want method PUT for PushCollectors, got", lastMethod)
-	}
-	if bytes.Compare(lastBody, wantBody) != 0 {
-		t.Errorf("got body %v, want %v", lastBody, wantBody)
-	}
-	if lastPath != "/metrics/job/testjob/instance/"+host {
-		t.Error("unexpected path:", lastPath)
-	}
-
-	// PushAddCollectors, with nil grouping, all good.
-	if err := AddCollectors("testjob", nil, pgwOK.URL, metric1, metric2); err != nil {
-		t.Fatal(err)
-	}
-	if lastMethod != "POST" {
-		t.Error("want method POST for PushAddCollectors, got", lastMethod)
+		t.Error("want method PUT for Push, got", lastMethod)
 	}
 	if bytes.Compare(lastBody, wantBody) != 0 {
 		t.Errorf("got body %v, want %v", lastBody, wantBody)
@@ -126,8 +103,28 @@ func TestPush(t *testing.T) {
 		t.Error("unexpected path:", lastPath)
 	}
 
-	// PushCollectors with a broken PGW.
-	if err := Collectors("testjob", nil, pgwErr.URL, metric1, metric2); err == nil {
+	// Add some Collectors, with nil grouping, all good.
+	if err := New(pgwOK.URL, "testjob").
+		Collector(metric1).
+		Collector(metric2).
+		Add(); err != nil {
+		t.Fatal(err)
+	}
+	if lastMethod != "POST" {
+		t.Error("want method POST for Add, got", lastMethod)
+	}
+	if bytes.Compare(lastBody, wantBody) != 0 {
+		t.Errorf("got body %v, want %v", lastBody, wantBody)
+	}
+	if lastPath != "/metrics/job/testjob" {
+		t.Error("unexpected path:", lastPath)
+	}
+
+	// Push some Collectors with a broken PGW.
+	if err := New(pgwErr.URL, "testjob").
+		Collector(metric1).
+		Collector(metric2).
+		Push(); err == nil {
 		t.Error("push to broken Pushgateway succeeded")
 	} else {
 		if got, want := err.Error(), "unexpected status code 500 while pushing to "+pgwErr.URL+"/metrics/job/testjob: fake error\n"; got != want {
@@ -135,22 +132,39 @@ func TestPush(t *testing.T) {
 		}
 	}
 
-	// PushCollectors with invalid grouping or job.
-	if err := Collectors("testjob", map[string]string{"foo": "bums"}, pgwErr.URL, metric1, metric2); err == nil {
+	// Push some Collectors with invalid grouping or job.
+	if err := New(pgwOK.URL, "testjob").
+		Grouping("foo", "bums").
+		Collector(metric1).
+		Collector(metric2).
+		Push(); err == nil {
 		t.Error("push with grouping contained in metrics succeeded")
 	}
-	if err := Collectors("test/job", nil, pgwErr.URL, metric1, metric2); err == nil {
+	if err := New(pgwOK.URL, "test/job").
+		Collector(metric1).
+		Collector(metric2).
+		Push(); err == nil {
 		t.Error("push with invalid job value succeeded")
 	}
-	if err := Collectors("testjob", map[string]string{"foo/bar": "bums"}, pgwErr.URL, metric1, metric2); err == nil {
+	if err := New(pgwOK.URL, "testjob").
+		Grouping("foobar", "bu/ms").
+		Collector(metric1).
+		Collector(metric2).
+		Push(); err == nil {
 		t.Error("push with invalid grouping succeeded")
 	}
-	if err := Collectors("testjob", map[string]string{"foo-bar": "bums"}, pgwErr.URL, metric1, metric2); err == nil {
+	if err := New(pgwOK.URL, "testjob").
+		Grouping("foo-bar", "bums").
+		Collector(metric1).
+		Collector(metric2).
+		Push(); err == nil {
 		t.Error("push with invalid grouping succeeded")
 	}
 
 	// Push registry, all good.
-	if err := FromGatherer("testjob", HostnameGroupingKey(), pgwOK.URL, reg); err != nil {
+	if err := New(pgwOK.URL, "testjob").
+		Gatherer(reg).
+		Push(); err != nil {
 		t.Fatal(err)
 	}
 	if lastMethod != "PUT" {
@@ -160,12 +174,16 @@ func TestPush(t *testing.T) {
 		t.Errorf("got body %v, want %v", lastBody, wantBody)
 	}
 
-	// PushAdd registry, all good.
-	if err := AddFromGatherer("testjob", map[string]string{"a": "x", "b": "y"}, pgwOK.URL, reg); err != nil {
+	// Add registry, all good.
+	if err := New(pgwOK.URL, "testjob").
+		Grouping("a", "x").
+		Grouping("b", "y").
+		Gatherer(reg).
+		Add(); err != nil {
 		t.Fatal(err)
 	}
 	if lastMethod != "POST" {
-		t.Error("want method POSTT for PushAdd, got", lastMethod)
+		t.Error("want method POST for Add, got", lastMethod)
 	}
 	if bytes.Compare(lastBody, wantBody) != 0 {
 		t.Errorf("got body %v, want %v", lastBody, wantBody)

vendor/github.com/prometheus/client_golang/prometheus/registry.go

@@ -18,6 +18,7 @@ import (
 	"errors"
 	"fmt"
 	"os"
+	"runtime"
 	"sort"
 	"sync"
 	"unicode/utf8"
@@ -36,13 +37,13 @@ const (
 // DefaultRegisterer and DefaultGatherer are the implementations of the
 // Registerer and Gatherer interface a number of convenience functions in this
 // package act on. Initially, both variables point to the same Registry, which
-// has a process collector (see NewProcessCollector) and a Go collector (see
-// NewGoCollector) already registered. This approach to keep default instances
-// as global state mirrors the approach of other packages in the Go standard
-// library. Note that there are caveats. Change the variables with caution and
-// only if you understand the consequences. Users who want to avoid global state
-// altogether should not use the convenience function and act on custom
-// instances instead.
+// has a process collector (currently on Linux only, see NewProcessCollector)
+// and a Go collector (see NewGoCollector) already registered. This approach to
+// keep default instances as global state mirrors the approach of other packages
+// in the Go standard library. Note that there are caveats. Change the variables
+// with caution and only if you understand the consequences. Users who want to
+// avoid global state altogether should not use the convenience functions and
+// act on custom instances instead.
 var (
 	defaultRegistry              = NewRegistry()
 	DefaultRegisterer Registerer = defaultRegistry
@@ -202,6 +203,13 @@ func (errs MultiError) Error() string {
 	return buf.String()
 }
 
+// Append appends the provided error if it is not nil.
+func (errs *MultiError) Append(err error) {
+	if err != nil {
+		*errs = append(*errs, err)
+	}
+}
+
 // MaybeUnwrap returns nil if len(errs) is 0. It returns the first and only
 // contained error as error if len(errs is 1). In all other cases, it returns
 // the MultiError directly. This is helpful for returning a MultiError in a way
@@ -368,22 +376,12 @@ func (r *Registry) Gather() ([]*dto.MetricFamily, error) {
 	)
 
 	r.mtx.RLock()
+	goroutineBudget := len(r.collectorsByID)
 	metricFamiliesByName := make(map[string]*dto.MetricFamily, len(r.dimHashesByName))
-
-	// Scatter.
-	// (Collectors could be complex and slow, so we call them all at once.)
-	wg.Add(len(r.collectorsByID))
-	go func() {
-		wg.Wait()
-		close(metricChan)
-	}()
+	collectors := make(chan Collector, len(r.collectorsByID))
 	for _, collector := range r.collectorsByID {
-		go func(collector Collector) {
-			defer wg.Done()
-			collector.Collect(metricChan)
-		}(collector)
+		collectors <- collector
 	}
-
 	// In case pedantic checks are enabled, we have to copy the map before
 	// giving up the RLock.
 	if r.pedanticChecksEnabled {
@@ -392,129 +390,176 @@ func (r *Registry) Gather() ([]*dto.MetricFamily, error) {
 			registeredDescIDs[id] = struct{}{}
 		}
 	}
-
 	r.mtx.RUnlock()
 
+	wg.Add(goroutineBudget)
+
+	collectWorker := func() {
+		for {
+			select {
+			case collector := <-collectors:
+				collector.Collect(metricChan)
+				wg.Done()
+			default:
+				return
+			}
+		}
+	}
+
+	// Start the first worker now to make sure at least one is running.
+	go collectWorker()
+	goroutineBudget--
+
+	// Close the metricChan once all collectors are collected.
+	go func() {
+		wg.Wait()
+		close(metricChan)
+	}()
+
 	// Drain metricChan in case of premature return.
 	defer func() {
 		for range metricChan {
 		}
 	}()
 
-	// Gather.
-	for metric := range metricChan {
-		// This could be done concurrently, too, but it required locking
-		// of metricFamiliesByName (and of metricHashes if checks are
-		// enabled). Most likely not worth it.
-		desc := metric.Desc()
-		dtoMetric := &dto.Metric{}
-		if err := metric.Write(dtoMetric); err != nil {
-			errs = append(errs, fmt.Errorf(
-				"error collecting metric %v: %s", desc, err,
+collectLoop:
+	for {
+		select {
+		case metric, ok := <-metricChan:
+			if !ok {
+				// metricChan is closed, we are done.
+				break collectLoop
+			}
+			errs.Append(processMetric(
+				metric, metricFamiliesByName,
+				metricHashes, dimHashes,
+				registeredDescIDs,
 			))
-			continue
+		default:
+			if goroutineBudget <= 0 || len(collectors) == 0 {
+				// All collectors are aleady being worked on or
+				// we have already as many goroutines started as
+				// there are collectors. Just process metrics
+				// from now on.
+				for metric := range metricChan {
+					errs.Append(processMetric(
+						metric, metricFamiliesByName,
+						metricHashes, dimHashes,
+						registeredDescIDs,
+					))
+				}
+				break collectLoop
+			}
+			// Start more workers.
+			go collectWorker()
+			goroutineBudget--
+			runtime.Gosched()
 		}
-		metricFamily, ok := metricFamiliesByName[desc.fqName]
-		if ok {
-			if metricFamily.GetHelp() != desc.help {
-				errs = append(errs, fmt.Errorf(
-					"collected metric %s %s has help %q but should have %q",
-					desc.fqName, dtoMetric, desc.help, metricFamily.GetHelp(),
-				))
-				continue
-			}
-			// TODO(beorn7): Simplify switch once Desc has type.
-			switch metricFamily.GetType() {
-			case dto.MetricType_COUNTER:
-				if dtoMetric.Counter == nil {
-					errs = append(errs, fmt.Errorf(
-						"collected metric %s %s should be a Counter",
-						desc.fqName, dtoMetric,
-					))
-					continue
-				}
-			case dto.MetricType_GAUGE:
-				if dtoMetric.Gauge == nil {
-					errs = append(errs, fmt.Errorf(
-						"collected metric %s %s should be a Gauge",
-						desc.fqName, dtoMetric,
-					))
-					continue
-				}
-			case dto.MetricType_SUMMARY:
-				if dtoMetric.Summary == nil {
-					errs = append(errs, fmt.Errorf(
-						"collected metric %s %s should be a Summary",
-						desc.fqName, dtoMetric,
-					))
-					continue
-				}
-			case dto.MetricType_UNTYPED:
-				if dtoMetric.Untyped == nil {
-					errs = append(errs, fmt.Errorf(
-						"collected metric %s %s should be Untyped",
-						desc.fqName, dtoMetric,
-					))
-					continue
-				}
-			case dto.MetricType_HISTOGRAM:
-				if dtoMetric.Histogram == nil {
-					errs = append(errs, fmt.Errorf(
-						"collected metric %s %s should be a Histogram",
-						desc.fqName, dtoMetric,
-					))
-					continue
-				}
-			default:
-				panic("encountered MetricFamily with invalid type")
-			}
-		} else {
-			metricFamily = &dto.MetricFamily{}
-			metricFamily.Name = proto.String(desc.fqName)
-			metricFamily.Help = proto.String(desc.help)
-			// TODO(beorn7): Simplify switch once Desc has type.
-			switch {
-			case dtoMetric.Gauge != nil:
-				metricFamily.Type = dto.MetricType_GAUGE.Enum()
-			case dtoMetric.Counter != nil:
-				metricFamily.Type = dto.MetricType_COUNTER.Enum()
-			case dtoMetric.Summary != nil:
-				metricFamily.Type = dto.MetricType_SUMMARY.Enum()
-			case dtoMetric.Untyped != nil:
-				metricFamily.Type = dto.MetricType_UNTYPED.Enum()
-			case dtoMetric.Histogram != nil:
-				metricFamily.Type = dto.MetricType_HISTOGRAM.Enum()
-			default:
-				errs = append(errs, fmt.Errorf(
-					"empty metric collected: %s", dtoMetric,
-				))
-				continue
-			}
-			metricFamiliesByName[desc.fqName] = metricFamily
-		}
-		if err := checkMetricConsistency(metricFamily, dtoMetric, metricHashes, dimHashes); err != nil {
-			errs = append(errs, err)
-			continue
-		}
-		if r.pedanticChecksEnabled {
-			// Is the desc registered at all?
-			if _, exist := registeredDescIDs[desc.id]; !exist {
-				errs = append(errs, fmt.Errorf(
-					"collected metric %s %s with unregistered descriptor %s",
-					metricFamily.GetName(), dtoMetric, desc,
-				))
-				continue
-			}
-			if err := checkDescConsistency(metricFamily, dtoMetric, desc); err != nil {
-				errs = append(errs, err)
-				continue
-			}
-		}
-		metricFamily.Metric = append(metricFamily.Metric, dtoMetric)
 	}
 	return normalizeMetricFamilies(metricFamiliesByName), errs.MaybeUnwrap()
 }
 
+// processMetric is an internal helper method only used by the Gather method.
+func processMetric(
+	metric Metric,
+	metricFamiliesByName map[string]*dto.MetricFamily,
+	metricHashes map[uint64]struct{},
+	dimHashes map[string]uint64,
+	registeredDescIDs map[uint64]struct{},
+) error {
+	desc := metric.Desc()
+	dtoMetric := &dto.Metric{}
+	if err := metric.Write(dtoMetric); err != nil {
+		return fmt.Errorf("error collecting metric %v: %s", desc, err)
+	}
+	metricFamily, ok := metricFamiliesByName[desc.fqName]
+	if ok {
+		if metricFamily.GetHelp() != desc.help {
+			return fmt.Errorf(
+				"collected metric %s %s has help %q but should have %q",
+				desc.fqName, dtoMetric, desc.help, metricFamily.GetHelp(),
+			)
+		}
+		// TODO(beorn7): Simplify switch once Desc has type.
+		switch metricFamily.GetType() {
+		case dto.MetricType_COUNTER:
+			if dtoMetric.Counter == nil {
+				return fmt.Errorf(
+					"collected metric %s %s should be a Counter",
+					desc.fqName, dtoMetric,
+				)
+			}
+		case dto.MetricType_GAUGE:
+			if dtoMetric.Gauge == nil {
+				return fmt.Errorf(
+					"collected metric %s %s should be a Gauge",
+					desc.fqName, dtoMetric,
+				)
+			}
+		case dto.MetricType_SUMMARY:
+			if dtoMetric.Summary == nil {
+				return fmt.Errorf(
+					"collected metric %s %s should be a Summary",
+					desc.fqName, dtoMetric,
+				)
+			}
+		case dto.MetricType_UNTYPED:
+			if dtoMetric.Untyped == nil {
+				return fmt.Errorf(
+					"collected metric %s %s should be Untyped",
+					desc.fqName, dtoMetric,
+				)
+			}
+		case dto.MetricType_HISTOGRAM:
+			if dtoMetric.Histogram == nil {
+				return fmt.Errorf(
+					"collected metric %s %s should be a Histogram",
+					desc.fqName, dtoMetric,
+				)
+			}
+		default:
+			panic("encountered MetricFamily with invalid type")
+		}
+	} else {
+		metricFamily = &dto.MetricFamily{}
+		metricFamily.Name = proto.String(desc.fqName)
+		metricFamily.Help = proto.String(desc.help)
+		// TODO(beorn7): Simplify switch once Desc has type.
+		switch {
+		case dtoMetric.Gauge != nil:
+			metricFamily.Type = dto.MetricType_GAUGE.Enum()
+		case dtoMetric.Counter != nil:
+			metricFamily.Type = dto.MetricType_COUNTER.Enum()
+		case dtoMetric.Summary != nil:
+			metricFamily.Type = dto.MetricType_SUMMARY.Enum()
+		case dtoMetric.Untyped != nil:
+			metricFamily.Type = dto.MetricType_UNTYPED.Enum()
+		case dtoMetric.Histogram != nil:
+			metricFamily.Type = dto.MetricType_HISTOGRAM.Enum()
+		default:
+			return fmt.Errorf("empty metric collected: %s", dtoMetric)
+		}
+		metricFamiliesByName[desc.fqName] = metricFamily
+	}
+	if err := checkMetricConsistency(metricFamily, dtoMetric, metricHashes, dimHashes); err != nil {
+		return err
+	}
+	if registeredDescIDs != nil {
+		// Is the desc registered at all?
+		if _, exist := registeredDescIDs[desc.id]; !exist {
+			return fmt.Errorf(
+				"collected metric %s %s with unregistered descriptor %s",
+				metricFamily.GetName(), dtoMetric, desc,
+			)
+		}
+		if err := checkDescConsistency(metricFamily, dtoMetric, desc); err != nil {
+			return err
+		}
+	}
+	metricFamily.Metric = append(metricFamily.Metric, dtoMetric)
+	return nil
+}
+
 // Gatherers is a slice of Gatherer instances that implements the Gatherer
 // interface itself. Its Gather method calls Gather on all Gatherers in the
 // slice in order and returns the merged results. Errors returned from the

vendor/github.com/prometheus/procfs/fs.go

@@ -46,8 +46,8 @@ func (fs FS) XFSStats() (*xfs.Stats, error) {
 	return xfs.ParseStats(f)
 }
 
-// NFSdClientRPCStats retrieves NFS daemon RPC statistics.
-func (fs FS) NFSdClientRPCStats() (*nfs.ClientRPCStats, error) {
+// NFSClientRPCStats retrieves NFS client RPC statistics.
+func (fs FS) NFSClientRPCStats() (*nfs.ClientRPCStats, error) {
 	f, err := os.Open(fs.Path("net/rpc/nfs"))
 	if err != nil {
 		return nil, err

vendor/github.com/prometheus/procfs/nfs/parse.go

@@ -178,8 +178,17 @@ func parseV3Stats(v []uint64) (V3Stats, error) {
 
 func parseClientV4Stats(v []uint64) (ClientV4Stats, error) {
 	values := int(v[0])
-	if len(v[1:]) != values || values < 59 {
-		return ClientV4Stats{}, fmt.Errorf("invalid V4Stats line %q", v)
+	if len(v[1:]) != values {
+		return ClientV4Stats{}, fmt.Errorf("invalid ClientV4Stats line %q", v)
+	}
+
+	// This function currently supports mapping 59 NFS v4 client stats.  Older
+	// kernels may emit fewer stats, so we must detect this and pad out the
+	// values to match the expected slice size.
+	if values < 59 {
+		newValues := make([]uint64, 60)
+		copy(newValues, v)
+		v = newValues
 	}
 
 	return ClientV4Stats{

vendor/github.com/prometheus/procfs/nfs/parse_nfs.go

@@ -32,12 +32,12 @@ func ParseClientRPCStats(r io.Reader) (*ClientRPCStats, error) {
 		parts := strings.Fields(scanner.Text())
 		// require at least <key> <value>
 		if len(parts) < 2 {
-			return nil, fmt.Errorf("invalid NFSd metric line %q", line)
+			return nil, fmt.Errorf("invalid NFS metric line %q", line)
 		}
 
 		values, err := util.ParseUint64s(parts[1:])
 		if err != nil {
-			return nil, fmt.Errorf("error parsing NFSd metric line: %s", err)
+			return nil, fmt.Errorf("error parsing NFS metric line: %s", err)
 		}
 
 		switch metricLine := parts[0]; metricLine {
@@ -52,15 +52,15 @@ func ParseClientRPCStats(r io.Reader) (*ClientRPCStats, error) {
 		case "proc4":
 			stats.ClientV4Stats, err = parseClientV4Stats(values)
 		default:
-			return nil, fmt.Errorf("unknown NFSd metric line %q", metricLine)
+			return nil, fmt.Errorf("unknown NFS metric line %q", metricLine)
 		}
 		if err != nil {
-			return nil, fmt.Errorf("errors parsing NFSd metric line: %s", err)
+			return nil, fmt.Errorf("errors parsing NFS metric line: %s", err)
 		}
 	}
 
 	if err := scanner.Err(); err != nil {
-		return nil, fmt.Errorf("error scanning NFSd file: %s", err)
+		return nil, fmt.Errorf("error scanning NFS file: %s", err)
 	}
 
 	return stats, nil

vendor/github.com/prometheus/procfs/nfs/parse_nfs_test.go

@@ -21,7 +21,7 @@ import (
 	"github.com/prometheus/procfs/nfs"
 )
 
-func TestNewNFSdClientRPCStats(t *testing.T) {
+func TestNewNFSClientRPCStats(t *testing.T) {
 	tests := []struct {
 		name    string
 		content string
@@ -32,6 +32,131 @@ func TestNewNFSdClientRPCStats(t *testing.T) {
 			name:    "invalid file",
 			content: "invalid",
 			invalid: true,
+		}, {
+			name: "good old kernel version file",
+			content: `net 70 70 69 45
+rpc 1218785755 374636 1218815394
+proc2 18 16 57 74 52 71 73 45 86 0 52 83 61 17 53 50 23 70 82
+proc3 22 0 1061909262 48906 4077635 117661341 5 29391916 2570425 2993289 590 0 0 7815 15 1130 0 3983 92385 13332 2 1 23729
+proc4 48 98 51 54 83 85 23 24 1 28 73 68 83 12 84 39 68 59 58 88 29 74 69 96 21 84 15 53 86 54 66 56 97 36 49 32 85 81 11 58 32 67 13 28 35 90 1 26 1337
+`,
+			stats: &nfs.ClientRPCStats{
+				Network: nfs.Network{
+					NetCount:   70,
+					UDPCount:   70,
+					TCPCount:   69,
+					TCPConnect: 45,
+				},
+				ClientRPC: nfs.ClientRPC{
+					RPCCount:        1218785755,
+					Retransmissions: 374636,
+					AuthRefreshes:   1218815394,
+				},
+				V2Stats: nfs.V2Stats{
+					Null:     16,
+					GetAttr:  57,
+					SetAttr:  74,
+					Root:     52,
+					Lookup:   71,
+					ReadLink: 73,
+					Read:     45,
+					WrCache:  86,
+					Write:    0,
+					Create:   52,
+					Remove:   83,
+					Rename:   61,
+					Link:     17,
+					SymLink:  53,
+					MkDir:    50,
+					RmDir:    23,
+					ReadDir:  70,
+					FsStat:   82,
+				},
+				V3Stats: nfs.V3Stats{
+					Null:        0,
+					GetAttr:     1061909262,
+					SetAttr:     48906,
+					Lookup:      4077635,
+					Access:      117661341,
+					ReadLink:    5,
+					Read:        29391916,
+					Write:       2570425,
+					Create:      2993289,
+					MkDir:       590,
+					SymLink:     0,
+					MkNod:       0,
+					Remove:      7815,
+					RmDir:       15,
+					Rename:      1130,
+					Link:        0,
+					ReadDir:     3983,
+					ReadDirPlus: 92385,
+					FsStat:      13332,
+					FsInfo:      2,
+					PathConf:    1,
+					Commit:      23729, },
+				ClientV4Stats: nfs.ClientV4Stats{
+					Null:               98,
+					Read:               51,
+					Write:              54,
+					Commit:             83,
+					Open:               85,
+					OpenConfirm:        23,
+					OpenNoattr:         24,
+					OpenDowngrade:      1,
+					Close:              28,
+					Setattr:            73,
+					FsInfo:             68,
+					Renew:              83,
+					SetClientId:        12,
+					SetClientIdConfirm: 84,
+					Lock:               39,
+					Lockt:              68,
+					Locku:              59,
+					Access:             58,
+					Getattr:            88,
+					Lookup:             29,
+					LookupRoot:         74,
+					Remove:             69,
+					Rename:             96,
+					Link:               21,
+					Symlink:            84,
+					Create:             15,
+					Pathconf:           53,
+					StatFs:             86,
+					ReadLink:           54,
+					ReadDir:            66,
+					ServerCaps:         56,
+					DelegReturn:        97,
+					GetAcl:             36,
+					SetAcl:             49,
+					FsLocations:        32,
+					ReleaseLockowner:   85,
+					Secinfo:            81,
+					FsidPresent:        11,
+					ExchangeId:         58,
+					CreateSession:      32,
+					DestroySession:     67,
+					Sequence:           13,
+					GetLeaseTime:       28,
+					ReclaimComplete:    35,
+					LayoutGet:          90,
+					GetDeviceInfo:      1,
+					LayoutCommit:       26,
+					LayoutReturn:       1337,
+					SecinfoNoName:      0,
+					TestStateId:        0,
+					FreeStateId:        0,
+					GetDeviceList:      0,
+					BindConnToSession:  0,
+					DestroyClientId:    0,
+					Seek:               0,
+					Allocate:           0,
+					DeAllocate:         0,
+					LayoutStats:        0,
+					Clone:              0,
+				},
+			},
 		}, {
 			name: "good file",
 			content: `net 18628 0 18628 6