Like Prometheus, but for logs.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
loki/pkg/storage/util_test.go

407 lines
12 KiB

package storage
import (
"context"
"sort"
"testing"
"time"
"github.com/davecgh/go-spew/spew"
"github.com/prometheus/common/model"
"github.com/prometheus/prometheus/model/labels"
"github.com/stretchr/testify/assert"
"github.com/grafana/loki/v3/pkg/chunkenc"
"github.com/grafana/loki/v3/pkg/ingester/client"
"github.com/grafana/loki/v3/pkg/logproto"
"github.com/grafana/loki/v3/pkg/logql/syntax"
"github.com/grafana/loki/v3/pkg/logqlmodel/stats"
"github.com/grafana/loki/v3/pkg/querier/astmapper"
"github.com/grafana/loki/v3/pkg/querier/plan"
"github.com/grafana/loki/v3/pkg/storage/chunk"
"github.com/grafana/loki/v3/pkg/storage/chunk/cache"
chunkclient "github.com/grafana/loki/v3/pkg/storage/chunk/client"
"github.com/grafana/loki/v3/pkg/storage/chunk/fetcher"
"github.com/grafana/loki/v3/pkg/storage/config"
"github.com/grafana/loki/v3/pkg/storage/stores"
index_stats "github.com/grafana/loki/v3/pkg/storage/stores/index/stats"
"github.com/grafana/loki/v3/pkg/storage/stores/shipper/indexshipper/tsdb/sharding"
loki_util "github.com/grafana/loki/v3/pkg/util"
"github.com/grafana/loki/v3/pkg/util/constants"
util_log "github.com/grafana/loki/v3/pkg/util/log"
)
var (
fooLabelsWithName = labels.Labels{{Name: "foo", Value: "bar"}, {Name: "__name__", Value: "logs"}}
fooLabels = labels.Labels{{Name: "foo", Value: "bar"}}
)
var from = time.Unix(0, time.Millisecond.Nanoseconds())
func assertStream(t *testing.T, expected, actual []logproto.Stream) {
if len(expected) != len(actual) {
t.Fatalf("error stream length are different expected %d actual %d\n%s", len(expected), len(actual), spew.Sdump(expected, actual))
return
}
sort.Slice(expected, func(i int, j int) bool { return expected[i].Labels < expected[j].Labels })
sort.Slice(actual, func(i int, j int) bool { return actual[i].Labels < actual[j].Labels })
for i := range expected {
assert.Equal(t, expected[i].Labels, actual[i].Labels)
if len(expected[i].Entries) != len(actual[i].Entries) {
t.Fatalf("error entries length are different expected %d actual %d\n%s", len(expected[i].Entries), len(actual[i].Entries), spew.Sdump(expected[i].Entries, actual[i].Entries))
return
}
for j := range expected[i].Entries {
assert.Equal(t, expected[i].Entries[j].Timestamp.UnixNano(), actual[i].Entries[j].Timestamp.UnixNano())
assert.Equal(t, expected[i].Entries[j].Line, actual[i].Entries[j].Line)
}
}
}
Improve metric queries by computing samples at the edges. (#2293) * First pass breaking the code appart. Wondering how we're going to achieve fast mutation of labels. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Work in progress. I realize I need hash for deduping lines. going to benchmark somes. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Tested some hash and decided which one to use. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Wip Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Starting working on ingester. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Trying to find a better hash function. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More hash testing we have a winner. xxhash it is. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Settle on xxhash Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Better params interfacing. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add interface for queryparams for things that exist in both type of params. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add storage sample iterator implementations. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing tests and verifying we don't get collions for the hashing method. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing ingesters tests and refactoring utility function/tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing and testing that stats are still well computed. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing more tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More engine tests finished. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes sharding evaluator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes more engine tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix error tests in the engine. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finish fixing all tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes a bug where extractor was not passed in correctly. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add notes about upgrade. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Renamed and fix a bug. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add memchunk tests and starting test for sampleIterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Test heap sample iterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * working on test. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finishing testing all new iterators. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Making sure all store functions are tested. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Benchmark and verify everything is working well. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Make the linter happy. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * use xxhash v2. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix a flaky test because of map. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * go.mod. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> Co-authored-by: Edward Welch <edward.welch@grafana.com>
6 years ago
func assertSeries(t *testing.T, expected, actual []logproto.Series) {
if len(expected) != len(actual) {
t.Fatalf("error stream length are different expected %d actual %d\n%s", len(expected), len(actual), spew.Sdump(expected, actual))
return
}
sort.Slice(expected, func(i int, j int) bool { return expected[i].Labels < expected[j].Labels })
sort.Slice(actual, func(i int, j int) bool { return actual[i].Labels < actual[j].Labels })
for i := range expected {
assert.Equal(t, expected[i].Labels, actual[i].Labels)
if len(expected[i].Samples) != len(actual[i].Samples) {
t.Fatalf("error entries length are different expected %d actual%d\n%s", len(expected[i].Samples), len(actual[i].Samples), spew.Sdump(expected[i].Samples, actual[i].Samples))
return
}
for j := range expected[i].Samples {
assert.Equal(t, expected[i].Samples[j].Timestamp, actual[i].Samples[j].Timestamp)
assert.Equal(t, expected[i].Samples[j].Value, actual[i].Samples[j].Value)
assert.Equal(t, expected[i].Samples[j].Hash, actual[i].Samples[j].Hash)
}
}
}
Pin `chunk` and `index` format to `schema` version. (#10213) We pin all three `Chunk`, `HeadBlock` and `TSDB` Version to `schema` version in period config. This is the following mapping (after being discussed with @owen-d and @sandeepsukhani ) `v12` (current existing schema) - ChunkFormatV3 (UnorderedHeadBlock) + TSDBv2 `v13` (introducing new schema) - ChunkFormatV4 (UnorderedWithNonIndexedLabelsHeadBlockFmt) + TSDBv3 Note the new schema `v13` supports the latest chunk and index format. **NOTES for Reviewer** 1. General approach is we removed the idea of `index.LiveFormat`, `chunkenc.DefaultChunkFormat` and `chunkenc.DefaultHeadBlockFmt` and made following two changes. These variables were used before to tie chunk and tsdb formats specific to Loki versions. This PR remove that coupling and pin these formats to `schema` version instead. 1. These variables were replaced with explicit chunk and index formats within those packages (and it's tests) 2. If these variables were used outside it's own packages say by ingester, compactor, etc. Then we extract correct chunk and index versions from the `schema` config. 2. Add two methods to `periodConfig`. (1) `ChunkFormat()` returning chunk and head format tied to schema (2) `TSDBFormat()` returning tsdb format tied to schema. 2. Other ideas I thought of doing but didn't end up doing is make `ChunkFormat` and `IndexFormat` as separate type (rather than `byte` and `int` currently. Similar to `HeadBlockFmt` type). But didnt' do it eventually to keep the PR small and don't want to complicate with lots of changes. 4. Moved couple of test cases from `chunkenc` to `config` package, because the test case was actually testing methods on `schemaconfig` and it was creating cycling dependencies. --------- Signed-off-by: Kaviraj <kavirajkanagaraj@gmail.com>
2 years ago
func newLazyChunk(chunkFormat byte, headfmt chunkenc.HeadBlockFmt, stream logproto.Stream) *LazyChunk {
return &LazyChunk{
Fetcher: nil,
IsValid: true,
Pin `chunk` and `index` format to `schema` version. (#10213) We pin all three `Chunk`, `HeadBlock` and `TSDB` Version to `schema` version in period config. This is the following mapping (after being discussed with @owen-d and @sandeepsukhani ) `v12` (current existing schema) - ChunkFormatV3 (UnorderedHeadBlock) + TSDBv2 `v13` (introducing new schema) - ChunkFormatV4 (UnorderedWithNonIndexedLabelsHeadBlockFmt) + TSDBv3 Note the new schema `v13` supports the latest chunk and index format. **NOTES for Reviewer** 1. General approach is we removed the idea of `index.LiveFormat`, `chunkenc.DefaultChunkFormat` and `chunkenc.DefaultHeadBlockFmt` and made following two changes. These variables were used before to tie chunk and tsdb formats specific to Loki versions. This PR remove that coupling and pin these formats to `schema` version instead. 1. These variables were replaced with explicit chunk and index formats within those packages (and it's tests) 2. If these variables were used outside it's own packages say by ingester, compactor, etc. Then we extract correct chunk and index versions from the `schema` config. 2. Add two methods to `periodConfig`. (1) `ChunkFormat()` returning chunk and head format tied to schema (2) `TSDBFormat()` returning tsdb format tied to schema. 2. Other ideas I thought of doing but didn't end up doing is make `ChunkFormat` and `IndexFormat` as separate type (rather than `byte` and `int` currently. Similar to `HeadBlockFmt` type). But didnt' do it eventually to keep the PR small and don't want to complicate with lots of changes. 4. Moved couple of test cases from `chunkenc` to `config` package, because the test case was actually testing methods on `schemaconfig` and it was creating cycling dependencies. --------- Signed-off-by: Kaviraj <kavirajkanagaraj@gmail.com>
2 years ago
Chunk: newChunk(chunkFormat, headfmt, stream),
}
}
Pin `chunk` and `index` format to `schema` version. (#10213) We pin all three `Chunk`, `HeadBlock` and `TSDB` Version to `schema` version in period config. This is the following mapping (after being discussed with @owen-d and @sandeepsukhani ) `v12` (current existing schema) - ChunkFormatV3 (UnorderedHeadBlock) + TSDBv2 `v13` (introducing new schema) - ChunkFormatV4 (UnorderedWithNonIndexedLabelsHeadBlockFmt) + TSDBv3 Note the new schema `v13` supports the latest chunk and index format. **NOTES for Reviewer** 1. General approach is we removed the idea of `index.LiveFormat`, `chunkenc.DefaultChunkFormat` and `chunkenc.DefaultHeadBlockFmt` and made following two changes. These variables were used before to tie chunk and tsdb formats specific to Loki versions. This PR remove that coupling and pin these formats to `schema` version instead. 1. These variables were replaced with explicit chunk and index formats within those packages (and it's tests) 2. If these variables were used outside it's own packages say by ingester, compactor, etc. Then we extract correct chunk and index versions from the `schema` config. 2. Add two methods to `periodConfig`. (1) `ChunkFormat()` returning chunk and head format tied to schema (2) `TSDBFormat()` returning tsdb format tied to schema. 2. Other ideas I thought of doing but didn't end up doing is make `ChunkFormat` and `IndexFormat` as separate type (rather than `byte` and `int` currently. Similar to `HeadBlockFmt` type). But didnt' do it eventually to keep the PR small and don't want to complicate with lots of changes. 4. Moved couple of test cases from `chunkenc` to `config` package, because the test case was actually testing methods on `schemaconfig` and it was creating cycling dependencies. --------- Signed-off-by: Kaviraj <kavirajkanagaraj@gmail.com>
2 years ago
func newLazyInvalidChunk(chunkFormat byte, headfmt chunkenc.HeadBlockFmt, stream logproto.Stream) *LazyChunk {
return &LazyChunk{
Fetcher: nil,
IsValid: false,
Pin `chunk` and `index` format to `schema` version. (#10213) We pin all three `Chunk`, `HeadBlock` and `TSDB` Version to `schema` version in period config. This is the following mapping (after being discussed with @owen-d and @sandeepsukhani ) `v12` (current existing schema) - ChunkFormatV3 (UnorderedHeadBlock) + TSDBv2 `v13` (introducing new schema) - ChunkFormatV4 (UnorderedWithNonIndexedLabelsHeadBlockFmt) + TSDBv3 Note the new schema `v13` supports the latest chunk and index format. **NOTES for Reviewer** 1. General approach is we removed the idea of `index.LiveFormat`, `chunkenc.DefaultChunkFormat` and `chunkenc.DefaultHeadBlockFmt` and made following two changes. These variables were used before to tie chunk and tsdb formats specific to Loki versions. This PR remove that coupling and pin these formats to `schema` version instead. 1. These variables were replaced with explicit chunk and index formats within those packages (and it's tests) 2. If these variables were used outside it's own packages say by ingester, compactor, etc. Then we extract correct chunk and index versions from the `schema` config. 2. Add two methods to `periodConfig`. (1) `ChunkFormat()` returning chunk and head format tied to schema (2) `TSDBFormat()` returning tsdb format tied to schema. 2. Other ideas I thought of doing but didn't end up doing is make `ChunkFormat` and `IndexFormat` as separate type (rather than `byte` and `int` currently. Similar to `HeadBlockFmt` type). But didnt' do it eventually to keep the PR small and don't want to complicate with lots of changes. 4. Moved couple of test cases from `chunkenc` to `config` package, because the test case was actually testing methods on `schemaconfig` and it was creating cycling dependencies. --------- Signed-off-by: Kaviraj <kavirajkanagaraj@gmail.com>
2 years ago
Chunk: newChunk(chunkFormat, headfmt, stream),
}
}
Pin `chunk` and `index` format to `schema` version. (#10213) We pin all three `Chunk`, `HeadBlock` and `TSDB` Version to `schema` version in period config. This is the following mapping (after being discussed with @owen-d and @sandeepsukhani ) `v12` (current existing schema) - ChunkFormatV3 (UnorderedHeadBlock) + TSDBv2 `v13` (introducing new schema) - ChunkFormatV4 (UnorderedWithNonIndexedLabelsHeadBlockFmt) + TSDBv3 Note the new schema `v13` supports the latest chunk and index format. **NOTES for Reviewer** 1. General approach is we removed the idea of `index.LiveFormat`, `chunkenc.DefaultChunkFormat` and `chunkenc.DefaultHeadBlockFmt` and made following two changes. These variables were used before to tie chunk and tsdb formats specific to Loki versions. This PR remove that coupling and pin these formats to `schema` version instead. 1. These variables were replaced with explicit chunk and index formats within those packages (and it's tests) 2. If these variables were used outside it's own packages say by ingester, compactor, etc. Then we extract correct chunk and index versions from the `schema` config. 2. Add two methods to `periodConfig`. (1) `ChunkFormat()` returning chunk and head format tied to schema (2) `TSDBFormat()` returning tsdb format tied to schema. 2. Other ideas I thought of doing but didn't end up doing is make `ChunkFormat` and `IndexFormat` as separate type (rather than `byte` and `int` currently. Similar to `HeadBlockFmt` type). But didnt' do it eventually to keep the PR small and don't want to complicate with lots of changes. 4. Moved couple of test cases from `chunkenc` to `config` package, because the test case was actually testing methods on `schemaconfig` and it was creating cycling dependencies. --------- Signed-off-by: Kaviraj <kavirajkanagaraj@gmail.com>
2 years ago
func newChunk(chunkFormat byte, headBlockFmt chunkenc.HeadBlockFmt, stream logproto.Stream) chunk.Chunk {
lbs, err := syntax.ParseLabels(stream.Labels)
if err != nil {
panic(err)
}
if !lbs.Has(labels.MetricName) {
builder := labels.NewBuilder(lbs)
builder.Set(labels.MetricName, "logs")
lbs = builder.Labels()
}
from, through := loki_util.RoundToMilliseconds(stream.Entries[0].Timestamp, stream.Entries[len(stream.Entries)-1].Timestamp)
Pin `chunk` and `index` format to `schema` version. (#10213) We pin all three `Chunk`, `HeadBlock` and `TSDB` Version to `schema` version in period config. This is the following mapping (after being discussed with @owen-d and @sandeepsukhani ) `v12` (current existing schema) - ChunkFormatV3 (UnorderedHeadBlock) + TSDBv2 `v13` (introducing new schema) - ChunkFormatV4 (UnorderedWithNonIndexedLabelsHeadBlockFmt) + TSDBv3 Note the new schema `v13` supports the latest chunk and index format. **NOTES for Reviewer** 1. General approach is we removed the idea of `index.LiveFormat`, `chunkenc.DefaultChunkFormat` and `chunkenc.DefaultHeadBlockFmt` and made following two changes. These variables were used before to tie chunk and tsdb formats specific to Loki versions. This PR remove that coupling and pin these formats to `schema` version instead. 1. These variables were replaced with explicit chunk and index formats within those packages (and it's tests) 2. If these variables were used outside it's own packages say by ingester, compactor, etc. Then we extract correct chunk and index versions from the `schema` config. 2. Add two methods to `periodConfig`. (1) `ChunkFormat()` returning chunk and head format tied to schema (2) `TSDBFormat()` returning tsdb format tied to schema. 2. Other ideas I thought of doing but didn't end up doing is make `ChunkFormat` and `IndexFormat` as separate type (rather than `byte` and `int` currently. Similar to `HeadBlockFmt` type). But didnt' do it eventually to keep the PR small and don't want to complicate with lots of changes. 4. Moved couple of test cases from `chunkenc` to `config` package, because the test case was actually testing methods on `schemaconfig` and it was creating cycling dependencies. --------- Signed-off-by: Kaviraj <kavirajkanagaraj@gmail.com>
2 years ago
chk := chunkenc.NewMemChunk(chunkFormat, chunkenc.EncGZIP, headBlockFmt, 256*1024, 0)
for _, e := range stream.Entries {
_ = chk.Append(&e)
}
chk.Close()
c := chunk.NewChunk("fake", client.Fingerprint(lbs), lbs, chunkenc.NewFacade(chk, 0, 0), from, through)
// force the checksum creation
if err := c.Encode(); err != nil {
panic(err)
}
return c
}
func newMatchers(matchers string) []*labels.Matcher {
res, err := syntax.ParseMatchers(matchers, true)
if err != nil {
panic(err)
}
return res
}
func newQuery(query string, start, end time.Time, shards []astmapper.ShardAnnotation, deletes []*logproto.Delete) *logproto.QueryRequest {
Feature/querysharding ii (#1927) * [wip] sharding evaluator/ast * [wip] continues experimenting with ast mapping * refactoring in preparation for binops * evaluators can pass state to other evaluators * compiler alignment * Evaluator method renamed to StepEvaluator * chained evaluator impl * tidying up sharding code * handling for ConcatSampleExpr * downstream iterator * structure for downstreaming asts * outlines sharding optimizations * work on sharding mapper * ast sharding optimizations * test for different logrange positions * shard mapper tests * stronger ast sharding & tests * shardmapper tests for string->string * removes sharding evaluator code * removes unused ctx arg * Revert "removes sharding evaluator code" This reverts commit 55d41b9519da9496e9471f13a5048d903ea04aaa. * interfaces for downstreaming, type conversions * sharding plumbing on frontend * type alignment in queryrange to downstream sharded queriers * downstreaming support for sharding incl storage code * removes chainedevaluator * comment alignment * storage shard injection * speccing out testware for sharding equivalence * [wip] shared engine refactor * sorting streams, sharding eval fixes * downstream evaluator embeds defaultevaluator * other pkgs adopt logql changes * metrics & logs use same middleware instantiation process * wires up shardingware * middleware per metrics/logfilter * empty step populating StepEvaluator promql.Matrix adapter * sharding metrics * log/span injection into sharded engine * sharding metrics avoids multiple instantiation * downstreamhandler tracing * sharding parameterized libsonnet * removes querier replicas * default 32 concurrency for workers * jsonnet correct level override * unquote true in yaml * lowercase error + downstreamEvaluator defaults to embedded defaultEvaluator * makes shardRecorder private * logs query on failed parse * refactors engine to be multi-use, minimizes logger injection, generalizes Query methods, removes Engine interface * basic tests for querysharding mware * [wip] concurrent evaluator * integrates stat propagation into sharding evaluator * splitby histogram * extends le bounds for bytes processed * byte throughput histogram buckets to 40gb * chunk duration mixin * fixes merge w/ field rename * derives logger in sharded engine via ctx & logs some downstream evaluators * moves sharded engine to top, adds comments * logs failed merge results in stats ctx * snapshotting stats merge logic is done more effectively * per query concurrency controlled via downstreamer * unexports decodereq * queryrange testware * downstreamer tests * pr requests
6 years ago
req := &logproto.QueryRequest{
Selector: query,
Start: start,
Limit: 1000,
End: end,
Improve metric queries by computing samples at the edges. (#2293) * First pass breaking the code appart. Wondering how we're going to achieve fast mutation of labels. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Work in progress. I realize I need hash for deduping lines. going to benchmark somes. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Tested some hash and decided which one to use. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Wip Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Starting working on ingester. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Trying to find a better hash function. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More hash testing we have a winner. xxhash it is. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Settle on xxhash Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Better params interfacing. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add interface for queryparams for things that exist in both type of params. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add storage sample iterator implementations. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing tests and verifying we don't get collions for the hashing method. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing ingesters tests and refactoring utility function/tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing and testing that stats are still well computed. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing more tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More engine tests finished. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes sharding evaluator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes more engine tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix error tests in the engine. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finish fixing all tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes a bug where extractor was not passed in correctly. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add notes about upgrade. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Renamed and fix a bug. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add memchunk tests and starting test for sampleIterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Test heap sample iterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * working on test. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finishing testing all new iterators. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Making sure all store functions are tested. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Benchmark and verify everything is working well. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Make the linter happy. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * use xxhash v2. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix a flaky test because of map. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * go.mod. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> Co-authored-by: Edward Welch <edward.welch@grafana.com>
6 years ago
Direction: logproto.FORWARD,
Deletes: deletes,
Plan: &plan.QueryPlan{
AST: syntax.MustParseExpr(query),
},
}
Feature/querysharding ii (#1927) * [wip] sharding evaluator/ast * [wip] continues experimenting with ast mapping * refactoring in preparation for binops * evaluators can pass state to other evaluators * compiler alignment * Evaluator method renamed to StepEvaluator * chained evaluator impl * tidying up sharding code * handling for ConcatSampleExpr * downstream iterator * structure for downstreaming asts * outlines sharding optimizations * work on sharding mapper * ast sharding optimizations * test for different logrange positions * shard mapper tests * stronger ast sharding & tests * shardmapper tests for string->string * removes sharding evaluator code * removes unused ctx arg * Revert "removes sharding evaluator code" This reverts commit 55d41b9519da9496e9471f13a5048d903ea04aaa. * interfaces for downstreaming, type conversions * sharding plumbing on frontend * type alignment in queryrange to downstream sharded queriers * downstreaming support for sharding incl storage code * removes chainedevaluator * comment alignment * storage shard injection * speccing out testware for sharding equivalence * [wip] shared engine refactor * sorting streams, sharding eval fixes * downstream evaluator embeds defaultevaluator * other pkgs adopt logql changes * metrics & logs use same middleware instantiation process * wires up shardingware * middleware per metrics/logfilter * empty step populating StepEvaluator promql.Matrix adapter * sharding metrics * log/span injection into sharded engine * sharding metrics avoids multiple instantiation * downstreamhandler tracing * sharding parameterized libsonnet * removes querier replicas * default 32 concurrency for workers * jsonnet correct level override * unquote true in yaml * lowercase error + downstreamEvaluator defaults to embedded defaultEvaluator * makes shardRecorder private * logs query on failed parse * refactors engine to be multi-use, minimizes logger injection, generalizes Query methods, removes Engine interface * basic tests for querysharding mware * [wip] concurrent evaluator * integrates stat propagation into sharding evaluator * splitby histogram * extends le bounds for bytes processed * byte throughput histogram buckets to 40gb * chunk duration mixin * fixes merge w/ field rename * derives logger in sharded engine via ctx & logs some downstream evaluators * moves sharded engine to top, adds comments * logs failed merge results in stats ctx * snapshotting stats merge logic is done more effectively * per query concurrency controlled via downstreamer * unexports decodereq * queryrange testware * downstreamer tests * pr requests
6 years ago
for _, shard := range shards {
req.Shards = append(req.Shards, shard.String())
}
return req
}
func newSampleQuery(query string, start, end time.Time, shards []astmapper.ShardAnnotation, deletes []*logproto.Delete) *logproto.SampleQueryRequest {
Improve metric queries by computing samples at the edges. (#2293) * First pass breaking the code appart. Wondering how we're going to achieve fast mutation of labels. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Work in progress. I realize I need hash for deduping lines. going to benchmark somes. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Tested some hash and decided which one to use. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Wip Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Starting working on ingester. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Trying to find a better hash function. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More hash testing we have a winner. xxhash it is. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Settle on xxhash Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Better params interfacing. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add interface for queryparams for things that exist in both type of params. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add storage sample iterator implementations. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing tests and verifying we don't get collions for the hashing method. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing ingesters tests and refactoring utility function/tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing and testing that stats are still well computed. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing more tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More engine tests finished. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes sharding evaluator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes more engine tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix error tests in the engine. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finish fixing all tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes a bug where extractor was not passed in correctly. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add notes about upgrade. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Renamed and fix a bug. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add memchunk tests and starting test for sampleIterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Test heap sample iterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * working on test. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finishing testing all new iterators. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Making sure all store functions are tested. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Benchmark and verify everything is working well. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Make the linter happy. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * use xxhash v2. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix a flaky test because of map. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * go.mod. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> Co-authored-by: Edward Welch <edward.welch@grafana.com>
6 years ago
req := &logproto.SampleQueryRequest{
Selector: query,
Start: start,
End: end,
Deletes: deletes,
Plan: &plan.QueryPlan{
AST: syntax.MustParseExpr(query),
},
Improve metric queries by computing samples at the edges. (#2293) * First pass breaking the code appart. Wondering how we're going to achieve fast mutation of labels. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Work in progress. I realize I need hash for deduping lines. going to benchmark somes. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Tested some hash and decided which one to use. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Wip Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Starting working on ingester. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Trying to find a better hash function. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More hash testing we have a winner. xxhash it is. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Settle on xxhash Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Better params interfacing. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add interface for queryparams for things that exist in both type of params. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add storage sample iterator implementations. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing tests and verifying we don't get collions for the hashing method. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing ingesters tests and refactoring utility function/tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing and testing that stats are still well computed. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing more tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More engine tests finished. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes sharding evaluator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes more engine tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix error tests in the engine. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finish fixing all tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes a bug where extractor was not passed in correctly. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add notes about upgrade. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Renamed and fix a bug. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add memchunk tests and starting test for sampleIterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Test heap sample iterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * working on test. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finishing testing all new iterators. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Making sure all store functions are tested. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Benchmark and verify everything is working well. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Make the linter happy. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * use xxhash v2. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix a flaky test because of map. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * go.mod. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> Co-authored-by: Edward Welch <edward.welch@grafana.com>
6 years ago
}
for _, shard := range shards {
req.Shards = append(req.Shards, shard.String())
}
Improve metric queries by computing samples at the edges. (#2293) * First pass breaking the code appart. Wondering how we're going to achieve fast mutation of labels. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Work in progress. I realize I need hash for deduping lines. going to benchmark somes. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Tested some hash and decided which one to use. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Wip Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Starting working on ingester. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Trying to find a better hash function. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More hash testing we have a winner. xxhash it is. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Settle on xxhash Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Better params interfacing. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add interface for queryparams for things that exist in both type of params. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add storage sample iterator implementations. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing tests and verifying we don't get collions for the hashing method. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing ingesters tests and refactoring utility function/tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing and testing that stats are still well computed. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixing more tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * More engine tests finished. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes sharding evaluator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes more engine tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix error tests in the engine. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finish fixing all tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes a bug where extractor was not passed in correctly. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add notes about upgrade. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Renamed and fix a bug. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add memchunk tests and starting test for sampleIterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Test heap sample iterator. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * working on test. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Finishing testing all new iterators. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Making sure all store functions are tested. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Benchmark and verify everything is working well. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Make the linter happy. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * use xxhash v2. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fix a flaky test because of map. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * go.mod. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> Co-authored-by: Edward Welch <edward.welch@grafana.com>
6 years ago
return req
}
type mockChunkStore struct {
schemas config.SchemaConfig
Simpler new chunk key v12 (#5054) * starts hacking new chunk paths * Create new chunk key for S3; update parsing and add basic test and benchmark Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> Fix `TestGrpcStore` to work with new chunk key structure Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> Add a function to SchemaConfig that returns the correct PeriodConfig for a given time. Signed-off-by: Callum Styan <callumstyan@gmail.com> Add schema config to object client so it we can use the right external key function based on the schema version. Signed-off-by: Callum Styan <callumstyan@gmail.com> Fix failing compactor tests by passing SchemaConfig to chunkClient Remove empty conditional from SchemaForTime Define schema v12; wire-in ChunkPathShardFactor and ChunkPathPeriod as configurable values Add PeriodConfig test steps for new values ChunkPathShardFactor and ChunkPathPeriod in v12 Update test for chunk.NewExternalKey(); remove completed TODO Set defaults for new ChunkPathPeriod and ChunkPathShardFactor SchemaConfig values; change FSObjectClient to use IdentityEncoder instead of Base64Encoder Use IdentityEncoder everywhere we use FSObjectClient for Chunks Add ExternalKey() function to SchemaConfig; update ObjectClient for Chunks * Finish plumbing through the chunk.ExternalKey -> schemaConfig.ExternalKey change. Signed-off-by: Callum Styan <callumstyan@gmail.com> * Clean up lint failures Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Clean up chunk.go and fix tests Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Break SchemaConfig.ExternalKey() into smaller functions Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Quickly fix variable name in SchemaConfig.newerExternalKey() Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Clean up ExternalKey conditional logic; add better comments; add tests Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Fix a bug where we are prepending userID to legacy Chunk keys but never parsing it; refactor key tests and benchmarks Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Add small SchemaConfig test Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Update docs and CHANGELOG.md Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Correctly return an error when failing to parse a chunk external key Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Revert IdentityEncoder to Base64Encoder after testing Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Fix assignment in test for linter Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Remove leftover comments from development Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Change v12 version comment format; remove redundant login in `ParseExternalKey()` Co-authored-by: Owen Diehl <ow.diehl@gmail.com> * Change remaining v12 comment style Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Pass chunk.SchemaConfig to new parallel chunk client * Simplify chunk external key prefixes for schema v12 * Fix broken benchmark; add benchmarks for old parsing conditional Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * Remove unneeded lines from upgrading doc * Add benchmarks for root chunk external key parsing functions Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> * memoizes schema number calculations * Resolve linter issue with PeriodConfig receiver name Signed-off-by: Jordan Rushing <jordan.rushing@grafana.com> Co-authored-by: Owen Diehl <ow.diehl@gmail.com> Co-authored-by: Callum Styan <callumstyan@gmail.com>
4 years ago
chunks []chunk.Chunk
client *mockChunkStoreClient
f chunk.RequestChunkFilterer
}
// mockChunkStore cannot implement both chunk.Store and chunk.Client,
// since there is a conflict in signature for DeleteChunk method.
var (
_ stores.Store = &mockChunkStore{}
_ chunkclient.Client = &mockChunkStoreClient{}
)
Pin `chunk` and `index` format to `schema` version. (#10213) We pin all three `Chunk`, `HeadBlock` and `TSDB` Version to `schema` version in period config. This is the following mapping (after being discussed with @owen-d and @sandeepsukhani ) `v12` (current existing schema) - ChunkFormatV3 (UnorderedHeadBlock) + TSDBv2 `v13` (introducing new schema) - ChunkFormatV4 (UnorderedWithNonIndexedLabelsHeadBlockFmt) + TSDBv3 Note the new schema `v13` supports the latest chunk and index format. **NOTES for Reviewer** 1. General approach is we removed the idea of `index.LiveFormat`, `chunkenc.DefaultChunkFormat` and `chunkenc.DefaultHeadBlockFmt` and made following two changes. These variables were used before to tie chunk and tsdb formats specific to Loki versions. This PR remove that coupling and pin these formats to `schema` version instead. 1. These variables were replaced with explicit chunk and index formats within those packages (and it's tests) 2. If these variables were used outside it's own packages say by ingester, compactor, etc. Then we extract correct chunk and index versions from the `schema` config. 2. Add two methods to `periodConfig`. (1) `ChunkFormat()` returning chunk and head format tied to schema (2) `TSDBFormat()` returning tsdb format tied to schema. 2. Other ideas I thought of doing but didn't end up doing is make `ChunkFormat` and `IndexFormat` as separate type (rather than `byte` and `int` currently. Similar to `HeadBlockFmt` type). But didnt' do it eventually to keep the PR small and don't want to complicate with lots of changes. 4. Moved couple of test cases from `chunkenc` to `config` package, because the test case was actually testing methods on `schemaconfig` and it was creating cycling dependencies. --------- Signed-off-by: Kaviraj <kavirajkanagaraj@gmail.com>
2 years ago
func newMockChunkStore(chunkFormat byte, headfmt chunkenc.HeadBlockFmt, streams []*logproto.Stream) *mockChunkStore {
chunks := make([]chunk.Chunk, 0, len(streams))
for _, s := range streams {
Pin `chunk` and `index` format to `schema` version. (#10213) We pin all three `Chunk`, `HeadBlock` and `TSDB` Version to `schema` version in period config. This is the following mapping (after being discussed with @owen-d and @sandeepsukhani ) `v12` (current existing schema) - ChunkFormatV3 (UnorderedHeadBlock) + TSDBv2 `v13` (introducing new schema) - ChunkFormatV4 (UnorderedWithNonIndexedLabelsHeadBlockFmt) + TSDBv3 Note the new schema `v13` supports the latest chunk and index format. **NOTES for Reviewer** 1. General approach is we removed the idea of `index.LiveFormat`, `chunkenc.DefaultChunkFormat` and `chunkenc.DefaultHeadBlockFmt` and made following two changes. These variables were used before to tie chunk and tsdb formats specific to Loki versions. This PR remove that coupling and pin these formats to `schema` version instead. 1. These variables were replaced with explicit chunk and index formats within those packages (and it's tests) 2. If these variables were used outside it's own packages say by ingester, compactor, etc. Then we extract correct chunk and index versions from the `schema` config. 2. Add two methods to `periodConfig`. (1) `ChunkFormat()` returning chunk and head format tied to schema (2) `TSDBFormat()` returning tsdb format tied to schema. 2. Other ideas I thought of doing but didn't end up doing is make `ChunkFormat` and `IndexFormat` as separate type (rather than `byte` and `int` currently. Similar to `HeadBlockFmt` type). But didnt' do it eventually to keep the PR small and don't want to complicate with lots of changes. 4. Moved couple of test cases from `chunkenc` to `config` package, because the test case was actually testing methods on `schemaconfig` and it was creating cycling dependencies. --------- Signed-off-by: Kaviraj <kavirajkanagaraj@gmail.com>
2 years ago
chunks = append(chunks, newChunk(chunkFormat, headfmt, *s))
}
return &mockChunkStore{schemas: config.SchemaConfig{}, chunks: chunks, client: &mockChunkStoreClient{chunks: chunks, scfg: config.SchemaConfig{}}}
}
func (m *mockChunkStore) Put(_ context.Context, _ []chunk.Chunk) error { return nil }
func (m *mockChunkStore) PutOne(_ context.Context, _, _ model.Time, _ chunk.Chunk) error {
return nil
}
func (m *mockChunkStore) GetSeries(ctx context.Context, _ string, _, _ model.Time, matchers ...*labels.Matcher) ([]labels.Labels, error) {
result := make([]labels.Labels, 0, len(m.chunks))
unique := map[uint64]struct{}{}
Outer:
for _, c := range m.chunks {
if _, ok := unique[c.Fingerprint]; !ok {
for _, m := range matchers {
if !m.Matches(c.Metric.Get(m.Name)) {
continue Outer
}
}
l := labels.NewBuilder(c.Metric).Del(labels.MetricName).Labels()
if m.f != nil {
if m.f.ForRequest(ctx).ShouldFilter(l) {
continue
}
}
result = append(result, l)
unique[c.Fingerprint] = struct{}{}
}
}
sort.Slice(result, func(i, j int) bool { return labels.Compare(result[i], result[j]) < 0 })
return result, nil
}
func (m *mockChunkStore) LabelValuesForMetricName(_ context.Context, _ string, _, _ model.Time, _ string, _ string, _ ...*labels.Matcher) ([]string, error) {
return nil, nil
}
func (m *mockChunkStore) LabelNamesForMetricName(_ context.Context, _ string, _, _ model.Time, _ string) ([]string, error) {
return nil, nil
}
func (m *mockChunkStore) SetChunkFilterer(f chunk.RequestChunkFilterer) {
m.f = f
}
func (m *mockChunkStore) DeleteChunk(_ context.Context, _, _ model.Time, _, _ string, _ labels.Labels, _ *model.Interval) error {
return nil
}
func (m *mockChunkStore) DeleteSeriesIDs(_ context.Context, _, _ model.Time, _ string, _ labels.Labels) error {
return nil
}
func (m *mockChunkStore) Stop() {}
func (m *mockChunkStore) Get(_ context.Context, _ string, _, _ model.Time, _ ...*labels.Matcher) ([]chunk.Chunk, error) {
return nil, nil
}
func (m *mockChunkStore) GetChunkFetcher(_ model.Time) *fetcher.Fetcher {
return nil
}
func (m *mockChunkStore) GetChunks(_ context.Context, _ string, _, _ model.Time, _ chunk.Predicate) ([][]chunk.Chunk, []*fetcher.Fetcher, error) {
refs := make([]chunk.Chunk, 0, len(m.chunks))
// transform real chunks into ref chunks.
for _, c := range m.chunks {
r, err := chunk.ParseExternalKey("fake", m.schemas.ExternalKey(c.ChunkRef))
if err != nil {
panic(err)
}
refs = append(refs, r)
}
cache, err := cache.New(cache.Config{Prefix: "chunks"}, nil, util_log.Logger, stats.ChunkCache, constants.Loki)
if err != nil {
panic(err)
}
f, err := fetcher.New(cache, nil, false, m.schemas, m.client, 0)
if err != nil {
panic(err)
}
return [][]chunk.Chunk{refs}, []*fetcher.Fetcher{f}, nil
}
func (m *mockChunkStore) Stats(_ context.Context, _ string, _, _ model.Time, _ ...*labels.Matcher) (*index_stats.Stats, error) {
return nil, nil
}
func (m *mockChunkStore) GetShards(_ context.Context, _ string, _, _ model.Time, _ uint64, _ chunk.Predicate) (*logproto.ShardsResponse, error) {
return nil, nil
}
func (m *mockChunkStore) HasForSeries(_, _ model.Time) (sharding.ForSeries, bool) {
return nil, false
}
func (m *mockChunkStore) Volume(_ context.Context, _ string, _, _ model.Time, _ int32, _ []string, _ string, _ ...*labels.Matcher) (*logproto.VolumeResponse, error) {
return nil, nil
}
type mockChunkStoreClient struct {
chunks []chunk.Chunk
scfg config.SchemaConfig
}
func (m mockChunkStoreClient) Stop() {
panic("implement me")
}
func (m mockChunkStoreClient) PutChunks(_ context.Context, _ []chunk.Chunk) error {
return nil
}
func (m mockChunkStoreClient) GetChunks(_ context.Context, chunks []chunk.Chunk) ([]chunk.Chunk, error) {
var res []chunk.Chunk
for _, c := range chunks {
for _, sc := range m.chunks {
// only returns chunks requested using the external key
if m.scfg.ExternalKey(c.ChunkRef) == m.scfg.ExternalKey(sc.ChunkRef) {
res = append(res, sc)
}
}
}
return res, nil
}
func (m mockChunkStoreClient) DeleteChunk(_ context.Context, _, _ string) error {
return nil
}
func (m mockChunkStoreClient) IsChunkNotFoundErr(_ error) bool {
return false
}
Dynamic client-side throttling to avoid object storage rate-limits (GCS only) (#10140) **What this PR does / why we need it**: Across the various cloud providers' object storage services, there are different rate-limits implemented. Rate-limits can be imposed under multiple conditions, such as server-side scale up (ramping up from low volume to high, "local" limit), reaching some defined upper limit ("absolute" limit), etc. We cannot know apriori when these rate-limits will be imposed, so we can't set up a client-side limiter to only allow a certain number of requests through per second. Additionally, that would require global coordination between queriers - which is difficult. With the above constraints, I have instead taken inspiration from TCP's [congestion control algorithms](https://en.wikipedia.org/wiki/TCP_congestion_control). This PR implements [AIMD](https://en.wikipedia.org/wiki/Additive_increase/multiplicative_decrease) (Additive Increase, Multiplicative Decrease), which is used in the congestion _avoidance_ phase of congestion control. The default window size (object store requests per second) is 2000; in other words, we skip the "slow start" phase. The controller uses the Go [`rate.Limiter`](https://pkg.go.dev/golang.org/x/time/rate), which implements the token-bucket algorithm. To put it simply: - every successful request widens the window (per second client rate-limit) - every rate-limited response reduces the window size by a backoff factor (0.5 by default, so it will halve) - when the limit has been reached, the querier will be delayed from making further requests until tokens are available
2 years ago
func (m mockChunkStoreClient) IsRetryableErr(_ error) bool {
return false
}
var streamsFixture = []*logproto.Stream{
{
Labels: "{foo=\"bar\"}",
Entries: []logproto.Entry{
{
Timestamp: from,
Line: "1",
},
{
Timestamp: from.Add(time.Millisecond),
Line: "2",
},
{
Timestamp: from.Add(2 * time.Millisecond),
Line: "3",
},
},
},
{
Labels: "{foo=\"bar\"}",
Entries: []logproto.Entry{
{
Timestamp: from.Add(2 * time.Millisecond),
Line: "3",
},
{
Timestamp: from.Add(3 * time.Millisecond),
Line: "4",
},
{
Timestamp: from.Add(4 * time.Millisecond),
Line: "5",
},
{
Timestamp: from.Add(5 * time.Millisecond),
Line: "6",
},
},
},
{
Labels: "{foo=\"bazz\"}",
Entries: []logproto.Entry{
{
Timestamp: from,
Line: "1",
},
{
Timestamp: from.Add(time.Millisecond),
Line: "2",
},
{
Timestamp: from.Add(2 * time.Millisecond),
Line: "3",
},
},
},
{
Labels: "{foo=\"bazz\"}",
Entries: []logproto.Entry{
{
Timestamp: from.Add(2 * time.Millisecond),
Line: "3",
},
{
Timestamp: from.Add(3 * time.Millisecond),
Line: "4",
},
{
Timestamp: from.Add(4 * time.Millisecond),
Line: "5",
},
{
Timestamp: from.Add(5 * time.Millisecond),
Line: "6",
},
},
},
}
var storeFixture = newMockChunkStore(chunkenc.ChunkFormatV3, chunkenc.UnorderedWithStructuredMetadataHeadBlockFmt, streamsFixture)