openshift · simonpasquier · May 11, 2020 · Jun 4, 2020 · Jun 4, 2020 · Jun 5, 2020
diff --git a/.circleci/config.yml b/.circleci/config.yml
@@ -87,11 +87,11 @@ jobs:
       - setup_remote_docker
       - run: ./fuzzit.sh fuzzing
 
-  makefile_sync:
+  repo_sync:
     executor: golang
     steps:
     - checkout
-    - run: ./scripts/sync_makefiles.sh
+    - run: ./scripts/sync_repo_files.sh
 
 workflows:
   version: 2
@@ -143,7 +143,7 @@ workflows:
               only:
                 - master
     jobs:
-    - makefile_sync:
+    - repo_sync:
         context: org-context
     - fuzzit_fuzzing:
         context: org-context
diff --git a/.github/workflows/funcbench.yml b/.github/workflows/funcbench.yml
@@ -9,7 +9,6 @@ jobs:
       AUTH_FILE: ${{ secrets.TEST_INFRA_GKE_AUTH }}
       BRANCH: ${{ github.event.client_payload.BRANCH }}
       BENCH_FUNC_REGEX: ${{ github.event.client_payload.BENCH_FUNC_REGEX }}
-      CLUSTER_NAME: test-infra
       GITHUB_TOKEN: ${{ secrets.PROMBOT_TOKEN }}
       GITHUB_ORG: prometheus
       GITHUB_REPO: prometheus

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,3 +1,38 @@
+## 2.20.0-rc.0 / 2020-07-16
+
+* [CHANGE] promtool: Changed rule numbering from 0-based to 1-based when reporting rule errors. #7495
+* [CHANGE] Remote read: Added `prometheus_remote_storage_read_queries_total` counter and `prometheus_remote_storage_read_request_duration_seconds` histogram, removed `prometheus_remote_storage_remote_read_queries_total` counter.
+* [CHANGE] Remote write: Added buckets for longer durations to `prometheus_remote_storage_sent_batch_duration_seconds` histogram.
+* [CHANGE] TSDB: WAL compression is enabled by default. #7410
+* [FEATURE] PromQL: Added `group()` aggregator. #7480
+* [FEATURE] SD: Added Docker Swarm SD. #7420
+* [FEATURE] SD: Added DigitalOcean SD. #7407
+* [FEATURE] SD: Added Openstack config option to query alternative endpoints. #7494
+* [ENHANCEMENT] Configuration: Exit early on invalid config file and signal it with exit code 2. #7399
+* [ENHANCEMENT] PromQL: `without` is now a valid metric identifier. #7533
+* [ENHANCEMENT] PromQL: Optimized regex label matching for literals within the pattern or as prefix/suffix. #7453 #7503
+* [ENHANCEMENT] promtool: Added time range parameters for labels API in promtool. #7463
+* [ENHANCEMENT] Remote write: Include samples waiting in channel in pending samples metric. Log number of dropped samples on hard shutdown. #7335
+* [ENHANCEMENT] Scrape: Ingest synthetic scrape report metrics atomically with the corresponding scraped metrics. #7562
+* [ENHANCEMENT] SD: Reduce timeouts for Openstack SD. #7507
+* [ENHANCEMENT] SD: Use 10m timeout for Consul watches. #7423
+* [ENHANCEMENT] SD: Added AMI meta label for EC2 SD. #7386
+* [ENHANCEMENT] TSDB: Increment WAL corruption metric also on WAL corruption during checkpointing. #7491
+* [ENHANCEMENT] TSDB: Improved query performance for high-cardinality labels. #7448
+* [ENHANCEMENT] UI: Display dates as well as timestamps in status page. #7544
+* [ENHANCEMENT] UI: Improved scrolling when following hash-fragment links. #7456
+* [ENHANCEMENT] UI: React UI renders numbers in alerts in a more human-readable way. #7426
+* [BUGFIX] API: Fixed error status code in the query API. #7435 
+* [BUGFIX] PromQL: Fixed `avg` and `avg_over_time` for NaN, Inf, and float64 overflows. #7346
+* [BUGFIX] PromQL: Fixed off-by-one error in `histogram_quantile`. #7393
+* [BUGFIX] promtool: Support extended durations in rules unit tests. #6297
+* [BUGFIX] Scrape: Fix undercounting for `scrape_samples_post_metric_relabeling` in case of errors. #7342
+* [BUGFIX] TSDB: Avoid leaving behind empty files in `chunks_head`, causing startup failures. #7573
+* [BUGFIX] TSDB: Fixed race between compact (gc, populate) and head append causing unknown symbol error. #7560
+* [BUGFIX] TSDB: Fixed unknown symbol error during head compaction. #7526
+* [BUGFIX] TSDB: Fixed panic during TSDB metric registration. #7501
+* [BUGFIX] TSDB: Fixed `--limit` command line flag in `tsdb` tool. #7430
+
 ## 2.19.2 / 2020-06-26
 
 * [BUGFIX] Remote Write: Fix panic when reloading config with modified queue parameters. #7452
@@ -18,6 +53,9 @@
 * [BUGFIX] React UI: Don't null out data when clicking on the current tab. #7243
 * [BUGFIX] PromQL: Correctly track number of samples for a query. #7307
 * [BUGFIX] PromQL: Return NaN when histogram buckets have 0 observations. #7318
+
+## 2.18.2 / 2020-06-09
+
 * [BUGFIX] TSDB: Fix incorrect query results when using Prometheus with remote reads configured #7361
 
 ## 2.18.1 / 2020-05-07

diff --git a/code-of-conduct.md → CODE_OF_CONDUCT.md b/code-of-conduct.md → CODE_OF_CONDUCT.md
diff --git a/Dockerfile.ocp b/Dockerfile.ocp
@@ -1,4 +1,4 @@
-FROM registry.svc.ci.openshift.org/openshift/release:golang-1.13 AS builder
+FROM registry.svc.ci.openshift.org/openshift/release:golang-1.14 AS builder
 WORKDIR /go/src/github.com/prometheus/prometheus
 COPY . .
 # NOTE(spasquie): the 'build' target regenerates the ReactJS code and the Go

diff --git a/MAINTAINERS.md b/MAINTAINERS.md
@@ -7,7 +7,7 @@
 * `documentation`
   * `prometheus-mixin`: @beorn7
 * `storage`
-  * `remote`: @csmarchbanks, @cstyan
+  * `remote`: @csmarchbanks, @cstyan, @bwplotka
 * `tsdb`: @codesome, @krasi-georgiev
 * `web`
   * `ui`: @juliusv

diff --git a/NOTICE b/NOTICE
@@ -86,6 +86,11 @@ https://github.com/samuel/go-zookeeper
 Copyright (c) 2013, Samuel Stauffer <samuel@descolada.com>
 See https://github.com/samuel/go-zookeeper/blob/master/LICENSE for license details.
 
+Time series compression algorithm from Facebook's Gorilla paper
+https://github.com/dgryski/go-tsz
+Copyright (c) 2015,2016 Damian Gryski <damian@gryski.com>
+See https://github.com/dgryski/go-tsz/blob/master/LICENSE for license details.
+
 We also use code from a large number of npm packages. For details, see:
 - https://github.com/prometheus/prometheus/blob/master/web/ui/react-app/package.json
 - https://github.com/prometheus/prometheus/blob/master/web/ui/react-app/package-lock.json

diff --git a/RELEASE.md b/RELEASE.md
@@ -25,7 +25,9 @@ Release cadence of first pre-releases being cut is 6 weeks.
 | v2.18          | 2020-04-22                                 | Bartek Plotka (GitHub: @bwplotka)           |
 | v2.19          | 2020-06-03                                 | Ganesh Vernekar (GitHub: @codesome)         |
 | v2.20          | 2020-07-15                                 | Björn Rabenstein (GitHub: @beorn7)          |
-| v2.21          | 2020-08-26                                 | **searching for volunteer**                 |
+| v2.21          | 2020-08-26                                 | Julien Pivotto (GitHub: @roidelapluie)      |
+| v2.22          | 2020-10-07                                 | Frederic Branczyk (GitHub: @brancz)         |
+| v2.23          | 2020-11-18                                 | **searching for volunteer**                 |
 
 If you are interested in volunteering please create a pull request against the [prometheus/prometheus](https://github.com/prometheus/prometheus) repository and propose yourself for the release series of your choice.
 

diff --git a/VERSION b/VERSION
@@ -1 +1 @@
-2.19.2
+2.20.0-rc.0
diff --git a/cmd/prometheus/main.go b/cmd/prometheus/main.go
@@ -214,7 +214,7 @@ func main() {
 		Default("false").BoolVar(&cfg.tsdb.AllowOverlappingBlocks)
 
 	a.Flag("storage.tsdb.wal-compression", "Compress the tsdb WAL.").
-		Default("false").BoolVar(&cfg.tsdb.WALCompression)
+		Default("true").BoolVar(&cfg.tsdb.WALCompression)
 
 	a.Flag("storage.remote.flush-deadline", "How long to wait flushing sample on shutdown or config reload.").
 		Default("1m").PlaceHolder("<duration>").SetValue(&cfg.RemoteFlushDeadline)
@@ -278,6 +278,12 @@ func main() {
 		os.Exit(2)
 	}
 
+	// Throw error for invalid config before starting other components.
+	if _, err := config.LoadFile(cfg.configFile); err != nil {
+		level.Error(logger).Log("msg", fmt.Sprintf("Error loading config (--config.file=%s)", cfg.configFile), "err", err)
+		os.Exit(2)
+	}
+
 	cfg.web.ReadTimeout = time.Duration(cfg.webTimeout)
 	// Default -web.route-prefix to path of -web.external-url.
 	if cfg.web.RoutePrefix == "" {
@@ -373,7 +379,7 @@ func main() {
 
 		ruleManager = rules.NewManager(&rules.ManagerOptions{
 			Appendable:      fanoutStorage,
-			TSDB:            localStorage,
+			Queryable:       localStorage,
 			QueryFunc:       rules.EngineQueryFunc(queryEngine, fanoutStorage),
 			NotifyFunc:      sendAlerts(notifierManager, cfg.web.ExternalURL.String()),
 			Context:         ctxRule,
@@ -955,6 +961,14 @@ func (s *readyStorage) Querier(ctx context.Context, mint, maxt int64) (storage.Q
 	return nil, tsdb.ErrNotReady
 }
 
+// ChunkQuerier implements the Storage interface.
+func (s *readyStorage) ChunkQuerier(ctx context.Context, mint, maxt int64) (storage.ChunkQuerier, error) {
+	if x := s.get(); x != nil {
+		return x.ChunkQuerier(ctx, mint, maxt)
+	}
+	return nil, tsdb.ErrNotReady
+}
+
 // Appender implements the Storage interface.
 func (s *readyStorage) Appender() storage.Appender {
 	if x := s.get(); x != nil {

diff --git a/cmd/prometheus/main_test.go b/cmd/prometheus/main_test.go
@@ -111,7 +111,7 @@ func TestFailedStartupExitCode(t *testing.T) {
 	}
 
 	fakeInputFile := "fake-input-file"
-	expectedExitStatus := 1
+	expectedExitStatus := 2
 
 	prom := exec.Command(promPath, "-test.main", "--config.file="+fakeInputFile)
 	err := prom.Run()

diff --git a/cmd/promtool/main.go b/cmd/promtool/main.go
@@ -95,6 +95,8 @@ func main() {
 	queryLabelsCmd := queryCmd.Command("labels", "Run labels query.")
 	queryLabelsServer := queryLabelsCmd.Arg("server", "Prometheus server to query.").Required().URL()
 	queryLabelsName := queryLabelsCmd.Arg("name", "Label name to provide label values for.").Required().String()
+	queryLabelsBegin := queryLabelsCmd.Flag("start", "Start time (RFC3339 or Unix timestamp).").String()
+	queryLabelsEnd := queryLabelsCmd.Flag("end", "End time (RFC3339 or Unix timestamp).").String()
 
 	testCmd := app.Command("test", "Unit testing.")
 	testRulesCmd := testCmd.Command("rules", "Unit tests for rules.")
@@ -142,7 +144,7 @@ func main() {
 		os.Exit(debugAll(*debugAllServer))
 
 	case queryLabelsCmd.FullCommand():
-		os.Exit(QueryLabels(*queryLabelsServer, *queryLabelsName, p))
+		os.Exit(QueryLabels(*queryLabelsServer, *queryLabelsName, *queryLabelsBegin, *queryLabelsEnd, p))
 
 	case testRulesCmd.FullCommand():
 		os.Exit(RulesUnitTest(*testRulesFiles...))
@@ -495,30 +497,10 @@ func QuerySeries(url *url.URL, matchers []string, start, end string, p printer)
 		return 1
 	}
 
-	// TODO: clean up timestamps
-	var (
-		minTime = time.Now().Add(-9999 * time.Hour)
-		maxTime = time.Now().Add(9999 * time.Hour)
-	)
-
-	var stime, etime time.Time
-
-	if start == "" {
-		stime = minTime
-	} else {
-		stime, err = parseTime(start)
-		if err != nil {
-			fmt.Fprintln(os.Stderr, "error parsing start time:", err)
-		}
-	}
-
-	if end == "" {
-		etime = maxTime
-	} else {
-		etime, err = parseTime(end)
-		if err != nil {
-			fmt.Fprintln(os.Stderr, "error parsing end time:", err)
-		}
+	stime, etime, err := parseStartTimeAndEndTime(start, end)
+	if err != nil {
+		fmt.Fprintln(os.Stderr, err)
+		return 1
 	}
 
 	// Run query against client.
@@ -537,7 +519,7 @@ func QuerySeries(url *url.URL, matchers []string, start, end string, p printer)
 }
 
 // QueryLabels queries for label values against a Prometheus server.
-func QueryLabels(url *url.URL, name string, p printer) int {
+func QueryLabels(url *url.URL, name string, start, end string, p printer) int {
 	config := api.Config{
 		Address: url.String(),
 	}
@@ -549,10 +531,16 @@ func QueryLabels(url *url.URL, name string, p printer) int {
 		return 1
 	}
 
+	stime, etime, err := parseStartTimeAndEndTime(start, end)
+	if err != nil {
+		fmt.Fprintln(os.Stderr, err)
+		return 1
+	}
+
 	// Run query against client.
 	api := v1.NewAPI(c)
 	ctx, cancel := context.WithTimeout(context.Background(), 2*time.Minute)
-	val, warn, err := api.LabelValues(ctx, name)
+	val, warn, err := api.LabelValues(ctx, name, stime, etime)
 	cancel()
 
 	for _, v := range warn {
@@ -568,6 +556,32 @@ func QueryLabels(url *url.URL, name string, p printer) int {
 	return 0
 }
 
+func parseStartTimeAndEndTime(start, end string) (time.Time, time.Time, error) {
+	var (
+		minTime = time.Now().Add(-9999 * time.Hour)
+		maxTime = time.Now().Add(9999 * time.Hour)
+		err     error
+	)
+
+	stime := minTime
+	etime := maxTime
+
+	if start != "" {
+		stime, err = parseTime(start)
+		if err != nil {
+			return stime, etime, errors.Wrap(err, "error parsing start time")
+		}
+	}
+
+	if end != "" {
+		etime, err = parseTime(end)
+		if err != nil {
+			return stime, etime, errors.Wrap(err, "error parsing end time")
+		}
+	}
+	return stime, etime, nil
+}
+
 func parseTime(s string) (time.Time, error) {
 	if t, err := strconv.ParseFloat(s, 64); err == nil {
 		s, ns := math.Modf(t)

diff --git a/cmd/promtool/testdata/alerts.yml b/cmd/promtool/testdata/alerts.yml
@@ -0,0 +1,13 @@
+# This is the rules file.
+
+groups:
+  - name: example
+    rules:
+      - alert: InstanceDown
+        expr: up == 0
+        for: 5m
+        labels:
+          severity: page
+        annotations:
+          summary: "Instance {{ $labels.instance }} down"
+          description: "{{ $labels.instance }} of job {{ $labels.job }} has been down for more than 5 minutes."
diff --git a/cmd/promtool/testdata/unittest.yml b/cmd/promtool/testdata/unittest.yml
@@ -0,0 +1,21 @@
+rule_files:
+  - alerts.yml
+
+evaluation_interval: 1m
+
+tests:
+  - interval: 1m
+    input_series:
+      - series: 'up{job="prometheus", instance="localhost:9090"}'
+        values: "0+0x1440"
+    alert_rule_test:
+      - eval_time: 1d
+        alertname: InstanceDown
+        exp_alerts:
+          - exp_labels:
+              severity: page
+              instance: localhost:9090
+              job: prometheus
+            exp_annotations:
+              summary: "Instance localhost:9090 down"
+              description: "localhost:9090 of job prometheus has been down for more than 5 minutes."