[PLAT-104961] Upgrade thanos to main and v0.35.0

.circleci/config.yml

@@ -29,7 +29,7 @@ jobs:
             make install-tool-deps
       - go/save-cache
       - setup_remote_docker:
-          version: 20.10.12
+          version: docker24
       - run:
           name: Create Secret if PR is not forked
           # GCS integration tests are run only for author's PR that have write access, because these tests
@@ -82,7 +82,7 @@ jobs:
       - git-shallow-clone/checkout
       - go/mod-download-cached
       - setup_remote_docker:
-          version: 20.10.12
+          version: docker24
       - attach_workspace:
           at: .
       # Register qemu to support multi-arch.
@@ -104,7 +104,7 @@ jobs:
       - git-shallow-clone/checkout
       - go/mod-download-cached
       - setup_remote_docker:
-          version: 20.10.12
+          version: docker24
       - attach_workspace:
           at: .
       - run: make tarballs-release

.github/workflows/react.yml

@@ -29,4 +29,5 @@ jobs:
           restore-keys: |
             ${{ runner.os }}-node-
 
+      - run: CI=false make check-react-app
       - run: make react-app-test

CHANGELOG.md

@@ -12,10 +12,29 @@ We use *breaking :warning:* to mark changes that are not backward compatible (re
 
 ### Fixed
 
+- [#7083](https://github.com/thanos-io/thanos/pull/7083) Store Gateway: Fix lazy expanded postings with 0 length failed to be cached.
+- [#7080](https://github.com/thanos-io/thanos/pull/7080) Receive: race condition in handler Close() when stopped early
+- [#7132](https://github.com/thanos-io/thanos/pull/7132) Documentation: fix broken helm installation instruction
+- [#7134](https://github.com/thanos-io/thanos/pull/7134) Store, Compact: Revert the recursive block listing mechanism introduced in https://github.com/thanos-io/thanos/pull/6474 and use the same strategy as in 0.31. Introduce a `--block-discovery-strategy` flag to control the listing strategy so that a recursive lister can still be used if the tradeoff of slower but cheaper discovery is preferred.
+- [#7122](https://github.com/thanos-io/thanos/pull/7122) Store Gateway: Fix lazy expanded postings estimate base cardinality using posting group with remove keys.
+- [#7224](https://github.com/thanos-io/thanos/pull/7224) Query-frontend: Add Redis username to the client configuration.
+- [#7220](https://github.com/thanos-io/thanos/pull/7220) Store Gateway: Fix lazy expanded postings caching partial expanded postings and bug of estimating remove postings with non existent value. Added `PromQLSmith` based fuzz test to improve correctness.
+
 ### Added
 
+- [#7194](https://github.com/thanos-io/thanos/pull/7194) Downsample: retry objstore related errors
+- [#7105](https://github.com/thanos-io/thanos/pull/7105) Rule: add flag `--query.enable-x-functions` to allow usage of extended promql functions (xrate, xincrease, xdelta) in loaded rules
+- [#6867](https://github.com/thanos-io/thanos/pull/6867) Query UI: Tenant input box added to the Query UI, in order to be able to specify which tenant the query should use.
+- [#7175](https://github.com/thanos-io/thanos/pull/7175): Query: Add `--query.mode=distributed` which enables the new distributed mode of the Thanos query engine.
+- [#7199](https://github.com/thanos-io/thanos/pull/7199): Reloader: Add support for watching and decompressing Prometheus configuration directories
+- [#7200](https://github.com/thanos-io/thanos/pull/7175): Query: Add `--selector.relabel-config` and `--selector.relabel-config-file` flags which allows scoping the Querier to a subset of matched TSDBs.
+- [#7233](https://github.com/thanos-io/thanos/pull/7233): UI: Showing Block Size Stats
+
 ### Changed
 
+- [#7123](https://github.com/thanos-io/thanos/pull/7123) Rule: Change default Alertmanager API version to v2.
+- [#7223](https://github.com/thanos-io/thanos/pull/7223) Automatic detection of memory limits and configure GOMEMLIMIT to match.
+
 ### Removed
 
 ## [v0.34.1](https://github.com/thanos-io/thanos/tree/release-0.34) - 11.02.24
@@ -38,6 +57,7 @@ We use *breaking :warning:* to mark changes that are not backward compatible (re
 - [#6874](https://github.com/thanos-io/thanos/pull/6874) Sidecar: fix labels returned by 'api/v1/series' in presence of conflicting external and inner labels.
 - [#7009](https://github.com/thanos-io/thanos/pull/7009) Rule: Fix spacing error in URL.
 - [#7082](https://github.com/thanos-io/thanos/pull/7082) Stores: fix label values edge case when requesting external label values with matchers
+- [#7114](https://github.com/thanos-io/thanos/pull/7114) Stores: fix file path bug for minio v7.0.61
 
 ### Added
 
@@ -53,13 +73,16 @@ We use *breaking :warning:* to mark changes that are not backward compatible (re
 - [#6887](https://github.com/thanos-io/thanos/pull/6887) Query Frontend: *breaking :warning:* Add tenant label to relevant exported metrics. Note that this change may cause some pre-existing custom dashboard queries to be incorrect due to the added label.
 - [#7028](https://github.com/thanos-io/thanos/pull/7028) Query|Query Frontend: Add new `--query-frontend.enable-x-functions` flag to enable experimental extended functions.
 - [#6884](https://github.com/thanos-io/thanos/pull/6884) Tools: Add upload-block command to upload blocks to object storage.
+- [#7010](https://github.com/thanos-io/thanos/pull/7010) Cache: Added `set_async_circuit_breaker_*` to utilize the circuit breaker pattern for dynamically thresholding asynchronous set operations.
 
 ### Changed
 
 - [#6539](https://github.com/thanos-io/thanos/pull/6539) Store: *breaking :warning:* Changed `--sync-block-duration` default 3m to 15m.
 
 ### Removed
 
+- [#7014](https://github.com/thanos-io/thanos/pull/7014) *: *breaking :warning:* Removed experimental query pushdown feature to simplify query path. This feature has had high complexity for too little benefits. The responsibility for query pushdown will be moved to the distributed mode of the new 'thanos' promql engine.
+
 ## [v0.33.0](https://github.com/thanos-io/thanos/tree/release-0.33) - 18.12.2023
 
 ### Fixed

Dockerfile.multi-stage

@@ -15,9 +15,8 @@ COPY . $GOPATH/src/github.com/thanos-io/thanos
 RUN git update-index --refresh; make build
 
 # -----------------------------------------------------------------------------
-FROM alpine:3.15
 
-#FROM quay.io/prometheus/busybox@sha256:${BASE_DOCKER_SHA}
+FROM quay.io/prometheus/busybox@sha256:${BASE_DOCKER_SHA}
 LABEL maintainer="The Thanos Authors"
 
 COPY --from=builder /go/bin/thanos /bin/thanos

MAINTAINERS.md

@@ -5,7 +5,7 @@
 | Bartłomiej Płotka     | [email protected]        | `@bwplotka`              | [@bwplotka](https://github.com/bwplotka)           | Google              |
 | Frederic Branczyk     | [email protected]       | `@brancz`                | [@brancz](https://github.com/brancz)               | Polar Signals       |
 | Giedrius Statkevičius | [email protected]    | `@Giedrius Statkevičius` | [@GiedriusS](https://github.com/GiedriusS)         | Vinted              |
-| Kemal Akkoyun         | [email protected]        | `@kakkoyun`              | [@kakkoyun](https://github.com/kakkoyun)           | Polar Signals       |
+| Kemal Akkoyun         | [email protected]        | `@kakkoyun`              | [@kakkoyun](https://github.com/kakkoyun)           | Fal                 |
 | Lucas Servén Marín    | [email protected]         | `@squat`                 | [@squat](https://github.com/squat)                 | Red Hat             |
 | Prem Saraswat         | [email protected]        | `@Prem Saraswat`         | [@onprem](https://github.com/onprem)               | Red Hat             |
 | Matthias Loibl        | [email protected]    | `@metalmatze`            | [@metalmatze](https://github.com/metalmatze)       | Polar Signals       |

Makefile

@@ -121,6 +121,10 @@ $(REACT_APP_OUTPUT_DIR): $(REACT_APP_NODE_MODULES_PATH) $(REACT_APP_SOURCE_FILES
 .PHONY: react-app
 react-app: $(REACT_APP_OUTPUT_DIR)
 
+.PHONY: check-react-app
+check-react-app: react-app
+	$(call require_clean_work_tree,'all generated files should be committed, run make react-app and commit changes.')
+
 .PHONY: react-app-lint
 react-app-lint: $(REACT_APP_NODE_MODULES_PATH)
 	   @echo ">> running React app linting"

VERSION

@@ -1 +1 @@
-0.34.1
+0.35.0-dev

cmd/thanos/compact.go

@@ -239,8 +239,16 @@ func runCompact(
 	consistencyDelayMetaFilter := block.NewConsistencyDelayMetaFilter(logger, conf.consistencyDelay, extprom.WrapRegistererWithPrefix("thanos_", reg))
 	timePartitionMetaFilter := block.NewTimePartitionMetaFilter(conf.filterConf.MinTime, conf.filterConf.MaxTime)
 
-	baseBlockIDsFetcher := block.NewBaseBlockIDsFetcher(logger, insBkt)
-	baseMetaFetcher, err := block.NewBaseFetcher(logger, conf.blockMetaFetchConcurrency, insBkt, baseBlockIDsFetcher, conf.dataDir, extprom.WrapRegistererWithPrefix("thanos_", reg))
+	var blockLister block.Lister
+	switch syncStrategy(conf.blockListStrategy) {
+	case concurrentDiscovery:
+		blockLister = block.NewConcurrentLister(logger, insBkt)
+	case recursiveDiscovery:
+		blockLister = block.NewRecursiveLister(logger, insBkt)
+	default:
+		return errors.Errorf("unknown sync strategy %s", conf.blockListStrategy)
+	}
+	baseMetaFetcher, err := block.NewBaseFetcher(logger, conf.blockMetaFetchConcurrency, insBkt, blockLister, conf.dataDir, extprom.WrapRegistererWithPrefix("thanos_", reg))
 	if err != nil {
 		return errors.Wrap(err, "create meta fetcher")
 	}
@@ -695,6 +703,7 @@ type compactConfig struct {
 	wait                                           bool
 	waitInterval                                   time.Duration
 	disableDownsampling                            bool
+	blockListStrategy                              string
 	blockMetaFetchConcurrency                      int
 	blockFilesConcurrency                          int
 	blockViewerSyncBlockInterval                   time.Duration
@@ -757,6 +766,9 @@ func (cc *compactConfig) registerFlag(cmd extkingpin.FlagClause) {
 		"as querying long time ranges without non-downsampled data is not efficient and useful e.g it is not possible to render all samples for a human eye anyway").
 		Default("false").BoolVar(&cc.disableDownsampling)
 
+	strategies := strings.Join([]string{string(concurrentDiscovery), string(recursiveDiscovery)}, ", ")
+	cmd.Flag("block-discovery-strategy", "One of "+strategies+". When set to concurrent, stores will concurrently issue one call per directory to discover active blocks in the bucket. The recursive strategy iterates through all objects in the bucket, recursively traversing into each directory. This avoids N+1 calls at the expense of having slower bucket iterations.").
+		Default(string(concurrentDiscovery)).StringVar(&cc.blockListStrategy)
 	cmd.Flag("block-meta-fetch-concurrency", "Number of goroutines to use when fetching block metadata from object storage.").
 		Default("32").IntVar(&cc.blockMetaFetchConcurrency)
 	cmd.Flag("block-files-concurrency", "Number of goroutines to use when fetching/uploading block files from object storage.").

cmd/thanos/config.go

@@ -13,6 +13,7 @@ import (
 	"strings"
 	"time"
 
+	"github.com/KimMachineGun/automemlimit/memlimit"
 	extflag "github.com/efficientgo/tools/extkingpin"
 	"github.com/pkg/errors"
 
@@ -283,3 +284,42 @@ func parseFlagLabels(s []string) (labels.Labels, error) {
 	sort.Sort(lset)
 	return lset, nil
 }
+
+type goMemLimitConfig struct {
+	enableAutoGoMemlimit bool
+	memlimitRatio        float64
+}
+
+func (gml *goMemLimitConfig) registerFlag(cmd extkingpin.FlagClause) *goMemLimitConfig {
+	cmd.Flag("enable-auto-gomemlimit",
+		"Enable go runtime to automatically limit memory consumption.").
+		Default("false").BoolVar(&gml.enableAutoGoMemlimit)
+
+	cmd.Flag("auto-gomemlimit.ratio",
+		"The ratio of reserved GOMEMLIMIT memory to the detected maximum container or system memory.").
+		Default("0.9").FloatVar(&gml.memlimitRatio)
+
+	return gml
+}
+
+func configureGoAutoMemLimit(common goMemLimitConfig) error {
+	if common.memlimitRatio <= 0.0 || common.memlimitRatio > 1.0 {
+		return errors.New("--auto-gomemlimit.ratio must be greater than 0 and less than or equal to 1.")
+	}
+
+	if common.enableAutoGoMemlimit {
+		if _, err := memlimit.SetGoMemLimitWithOpts(
+			memlimit.WithRatio(common.memlimitRatio),
+			memlimit.WithProvider(
+				memlimit.ApplyFallback(
+					memlimit.FromCgroup,
+					memlimit.FromSystem,
+				),
+			),
+		); err != nil {
+			return errors.Wrap(err, "Failed to set GOMEMLIMIT automatically")
+		}
+	}
+
+	return nil
+}

cmd/thanos/downsample.go

@@ -21,6 +21,7 @@ import (
 	"github.com/prometheus/client_golang/prometheus/promauto"
 	"github.com/prometheus/prometheus/tsdb"
 	"github.com/prometheus/prometheus/tsdb/chunkenc"
+	"github.com/thanos-io/thanos/pkg/compact"
 
 	"github.com/thanos-io/objstore"
 	"github.com/thanos-io/objstore/client"
@@ -90,7 +91,7 @@ func RunDownsample(
 	insBkt := objstoretracing.WrapWithTraces(objstore.WrapWithMetrics(bkt, extprom.WrapRegistererWithPrefix("thanos_", reg), bkt.Name()))
 
 	// While fetching blocks, filter out blocks that were marked for no downsample.
-	baseBlockIDsFetcher := block.NewBaseBlockIDsFetcher(logger, insBkt)
+	baseBlockIDsFetcher := block.NewConcurrentLister(logger, insBkt)
 	metaFetcher, err := block.NewMetaFetcher(logger, block.FetcherConcurrency, insBkt, baseBlockIDsFetcher, "", extprom.WrapRegistererWithPrefix("thanos_", reg), []block.MetadataFilter{
 		block.NewDeduplicateFilter(block.FetcherConcurrency),
 		downsample.NewGatherNoDownsampleMarkFilter(logger, insBkt, block.FetcherConcurrency),
@@ -358,7 +359,7 @@ func processDownsampling(
 
 	err := block.Download(ctx, logger, bkt, m.ULID, bdir, objstore.WithFetchConcurrency(blockFilesConcurrency))
 	if err != nil {
-		return errors.Wrapf(err, "download block %s", m.ULID)
+		return compact.NewRetryError(errors.Wrapf(err, "download block %s", m.ULID))
 	}
 	level.Info(logger).Log("msg", "downloaded block", "id", m.ULID, "duration", time.Since(begin), "duration_ms", time.Since(begin).Milliseconds())
 
@@ -419,7 +420,7 @@ func processDownsampling(
 
 	err = block.Upload(ctx, logger, bkt, resdir, hashFunc)
 	if err != nil {
-		return errors.Wrapf(err, "upload downsampled block %s", id)
+		return compact.NewRetryError(errors.Wrapf(err, "upload downsampled block %s", id))
 	}
 
 	level.Info(logger).Log("msg", "uploaded block", "id", id, "duration", time.Since(begin), "duration_ms", time.Since(begin).Milliseconds())

cmd/thanos/main.go

@@ -22,6 +22,7 @@ import (
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/collectors"
+	versioncollector "github.com/prometheus/client_golang/prometheus/collectors/version"
 	"github.com/prometheus/common/version"
 	"go.uber.org/automaxprocs/maxprocs"
 	"gopkg.in/alecthomas/kingpin.v2"
@@ -49,6 +50,10 @@ func main() {
 		Default(logging.LogFormatLogfmt).Enum(logging.LogFormatLogfmt, logging.LogFormatJSON)
 	tracingConfig := extkingpin.RegisterCommonTracingFlags(app)
 
+	goMemLimitConf := goMemLimitConfig{}
+
+	goMemLimitConf.registerFlag(app)
+
 	registerSidecar(app)
 	registerStore(app)
 	registerQuery(app)
@@ -61,6 +66,11 @@ func main() {
 	cmd, setup := app.Parse()
 	logger := logging.NewLogger(*logLevel, *logFormat, *debugName)
 
+	if err := configureGoAutoMemLimit(goMemLimitConf); err != nil {
+		level.Error(logger).Log("msg", "failed to configure Go runtime memory limits", "err", err)
+		os.Exit(1)
+	}
+
 	// Running in container with limits but with empty/wrong value of GOMAXPROCS env var could lead to throttling by cpu
 	// maxprocs will automate adjustment by using cgroups info about cpu limit if it set as value for runtime.GOMAXPROCS.
 	undo, err := maxprocs.Set(maxprocs.Logger(func(template string, args ...interface{}) {
@@ -73,7 +83,7 @@ func main() {
 
 	metrics := prometheus.NewRegistry()
 	metrics.MustRegister(
-		version.NewCollector("thanos"),
+		versioncollector.NewCollector("thanos"),
 		collectors.NewGoCollector(
 			collectors.WithGoCollectorRuntimeMetrics(collectors.GoRuntimeMetricsRule{Matcher: regexp.MustCompile("/.*")}),
 		),

cmd/thanos/main_test.go

@@ -21,6 +21,7 @@ import (
 	"github.com/thanos-io/objstore"
 
 	"github.com/efficientgo/core/testutil"
+
 	"github.com/thanos-io/thanos/pkg/block"
 	"github.com/thanos-io/thanos/pkg/block/metadata"
 	"github.com/thanos-io/thanos/pkg/compact/downsample"
@@ -157,7 +158,7 @@ func TestRegression4960_Deadlock(t *testing.T) {
 
 	metrics := newDownsampleMetrics(prometheus.NewRegistry())
 	testutil.Equals(t, 0.0, promtest.ToFloat64(metrics.downsamples.WithLabelValues(meta.Thanos.GroupKey())))
-	baseBlockIDsFetcher := block.NewBaseBlockIDsFetcher(logger, bkt)
+	baseBlockIDsFetcher := block.NewConcurrentLister(logger, bkt)
 	metaFetcher, err := block.NewMetaFetcher(nil, block.FetcherConcurrency, bkt, baseBlockIDsFetcher, "", nil, nil)
 	testutil.Ok(t, err)
 
@@ -197,7 +198,7 @@ func TestCleanupDownsampleCacheFolder(t *testing.T) {
 
 	metrics := newDownsampleMetrics(prometheus.NewRegistry())
 	testutil.Equals(t, 0.0, promtest.ToFloat64(metrics.downsamples.WithLabelValues(meta.Thanos.GroupKey())))
-	baseBlockIDsFetcher := block.NewBaseBlockIDsFetcher(logger, bkt)
+	baseBlockIDsFetcher := block.NewConcurrentLister(logger, bkt)
 	metaFetcher, err := block.NewMetaFetcher(nil, block.FetcherConcurrency, bkt, baseBlockIDsFetcher, "", nil, nil)
 	testutil.Ok(t, err)