apitech/loki - loki - Gitea: Git with a cup of tea

Commit Graph

Author	SHA1	Message	Date
Salva Corts	d24fe3e68b	Max bytes read limit (#8670 ) What this PR does / why we need it: This PR implements two new per-tenant limits that are enforced on log and metric queries (both range and instant) when TSDB is used: - `max_query_bytes_read`: Refuse queries that would read more than the configured bytes here. Overall limit regardless of splitting/sharding. The goal is to refuse queries that would take too long. The default value of 0 disables this limit. - `max_querier_bytes_read`: Refuse queries in which any of their subqueries after splitting and sharding would read more than the configured bytes here. The goal is to avoid a querier from running a query that would load too much data in memory and can potentially get OOMed. The default value of 0 disables this limit. These new limits can be configured per tenant and per query (see https://github.com/grafana/loki/pull/8727). The bytes a query would read are estimated through TSDB's index stats. Even though they are not exact, they are good enough to have a rough estimation of whether a query is too big to run or not. For more details on this refer to this discussion in the PR: https://github.com/grafana/loki/pull/8670#discussion_r1124858508. Both limits are implemented in the frontend. Even though we considered implementing `max_querier_bytes_read` in the querier, this way, the limits for pre and post splitting/sharding queries are enforced close to each other on the same component. Moreover, this way we can reduce the number of index stats requests issued to the index gateways by reusing the stats gathered while sharding the query. With regard to how index stats requests are issued: - We parallelize index stats requests by splitting them into queries that span up to 24h since our indices are sharded by 24h periods. On top of that, this prevents a single index gateway from processing a single huge request like `{app=~".+"} for 30d`. - If sharding is enabled and the query is shardable, for `max_querier_bytes_read`, we re-use the stats requests issued by the sharding ware. Specifically, we look at the [bytesPerShard][1] to enforce this limit. Note that once we merge this PR and enable these limits, the load of index stats requests will increase substantially and we may discover bottlenecks in our index gateways and TSDB. After speaking with @owen-d, we think it should be fine as, if needed, we can scale up our index gateways and support caching index stats requests. Here's a demo of this working: <img width="1647" alt="image" src="https://user-images.githubusercontent.com/8354290/226918478-d4b6c2fd-de4d-478a-9c8b-e38fe148fa95.png"> <img width="1647" alt="image" src="https://user-images.githubusercontent.com/8354290/226918798-a71b1db8-ea68-4d00-933b-e5eb1524d240.png"> Which issue(s) this PR fixes: This PR addresses https://github.com/grafana/loki-private/issues/674. Special notes for your reviewer: - @jeschkies has reviewed the changes related to query-time limits. - I've done some refactoring in this PR: - Extracted logic to get stats for a set of matches into a new function [getStatsForMatchers][2]. - Extracted the _Handler_ interface implementation for [queryrangebase.roundTripper][3] into a new type [queryrangebase.roundTripperHandler][4]. This is used to create the handler that skips the rest of configured middlewares when sending an index stat quests ([example][5]). Checklist - [x] Reviewed the [`CONTRIBUTING.md`](https://github.com/grafana/loki/blob/main/CONTRIBUTING.md) guide (required) - [x] Documentation added - [x] Tests updated - [x] `CHANGELOG.md` updated - [ ] Changes that require user attention or interaction to upgrade are documented in `docs/sources/upgrading/_index.md` [1]: `ff847305af/pkg/querier/queryrange/shard_resolver.go (L179-L186)` [2]: `ff847305af/pkg/querier/queryrange/shard_resolver.go (L72)` [3]: `3d2fff3a2d/pkg/querier/queryrange/queryrangebase/roundtrip.go (L124)` [4]: `3d2fff3a2d/pkg/querier/queryrange/queryrangebase/roundtrip.go (L163)` [5]: `f422e0a52b/pkg/querier/queryrange/roundtrip.go (L521)`	3 years ago
Ed Welch	a4eb536fb2	Loki: remove `subqueries` from metrics.go logging and replace it with separate split and shard counters (#8761 ) What this PR does / why we need it: Currently the `metrics.go` log line emitted after every query includes a metric called "subqueries". This currently tracks the number of queries created by the split_by_time operations done in Loki, but does not include any counts for subqueries created as a result of sharding. It becomes difficult to make a single subqueries counter that gives useful information to determine how much a query is split by time and sharded by a shard factor, especially now that sharding in TSDB indexes is dynamic. This PR removes and deprecates the `subqueries` stat and instead creates a `splits` and `shards` statistic which records how much a query was split_by_time and the total number of shards created as well. Which issue(s) this PR fixes: Fixes #<issue number> Special notes for your reviewer: Checklist - [ ] Reviewed the [`CONTRIBUTING.md`](https://github.com/grafana/loki/blob/main/CONTRIBUTING.md) guide (required) - [ ] Documentation added - [x] Tests updated - [x] `CHANGELOG.md` updated - [ ] Changes that require user attention or interaction to upgrade are documented in `docs/sources/upgrading/_index.md` --------- Signed-off-by: Edward Welch <edward.welch@grafana.com>	3 years ago
Owen Diehl	22089415e8	Split parallelism across Period Configs (#7769 ) One of things we watch while updating non-TSDB period configs to TSDB period configs is the difference in query parallelism. TSDB dynamically shards queries into (potentially) much smaller units of work compared to the static shard factors uses prior. To account for this, we use much higher query parallelism configurations with TSDB period configs. This creates a potential problem when querying across `non-tsdb, tsdb` period boundaries: we may want a query parallelism of 512 for the tsdb portion but only 64 for the non-tsdb portion! However, we only had one limit to specify this per tenant, meaning this would be too high when querying non-tsdb periods or too low when querying tsdb ones. This PR * Introduces `tsdb_max_query_parallelism` (default `512`) to `limits_config` * Uses `tsdb_max_query_parallelism` and `max_query_parallelism` limits to find a better parallelism _per query_ by weighting the two respective configs by the proportion of each query spent on TSDB or non-TSDB period configurations. Signed-off-by: Owen Diehl <ow.diehl@gmail.com>	3 years ago
Christian Haudum	95b885691d	Preserve interval parameter when splitting queries by time (#5622 ) * Preserve interval parameter when splitting queries by time Signed-off-by: Christian Haudum <christian.haudum@gmail.com> * Add tests for interval param on Loki downstream requests Signed-off-by: Christian Haudum <christian.haudum@gmail.com> * Add changelog entry for #5622 Signed-off-by: Christian Haudum <christian.haudum@gmail.com>	4 years ago
Cyril Tovena	9b6f17f168	Add the total number of subqueries to query statistics (#5397 ) * Add subqueries count Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes tests Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com>	4 years ago
Owen Diehl	91d837e79c	Refactor/remove global splitby (#5243 ) * cleans up internal limits struct, removes unnecessary logic * moves querier.split-queries-by-interval to limits, simplifies resulting code. * docs, changelog * flag supports model.duration instead of time.duration * lint and adds new error for deprecated field * reregisters queryrange flags * comment lint * ensures we validate queryrange config and simplifies embedded struct validate() * 0 still disables * simplify cache key generation and protect against divide by zero	4 years ago
Callum Styan	f24b84c15a	[querier] Abort a query in limitedRoundTripper if parallelism is 0 (#5190 ) * Abort a query in limitedRoundTripper if parallelism is 0. Signed-off-by: Callum Styan <callumstyan@gmail.com> * Fix failing tests related to fakeLimits change. Signed-off-by: Callum Styan <callumstyan@gmail.com> * Change limitedRoundTripper parallelism error message. Signed-off-by: Callum Styan <callumstyan@gmail.com>	4 years ago
Sandeep Sukhani	cd1537d06e	step align start and end time of the original query while splitting it (#5217 ) * step align start and end time of the original query while splitting it * also step align queries with step > split interval	4 years ago
Sandeep Sukhani	dad7fca2fd	align metric queries by step and other queries by split interval (#5181 )	4 years ago
Cyril Tovena	1a7614ff70	Avoid splitting large range vector aggregation. (#5172 ) * Reduce split by for large range vector Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Update pkg/querier/queryrange/split_by_interval.go Co-authored-by: Christian Haudum <christian.haudum@gmail.com> * improve fn name. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> Co-authored-by: Christian Haudum <christian.haudum@gmail.com>	4 years ago
Dylan Guedes	0e3df1938f	Copy querier subpackages dependency into Loki (#5068 ) * Fork cortex `querier/series`. * Fork cortex `querier/chunkstore`. * Fork cortex `querier/lazyquery`. * Use Loki forked cortex code. * Copy cortex `querier/queryrange` as `querier/queryrange/queryrangebase`. * Fork cortex `pkg/querier/queryrange` to `pkg/querier/queryrange/queryrangebase`. * Tell protoc to use gogo.types instead of pb.Any. * Use `pkg/querier/queryrange/queryrangebase`. * Remove cortex `queryrange` vendoring. * Fix lint. * Use `_ = ...` syntax instead of comment to fix lint.	4 years ago
Sandeep Sukhani	b865afc0d7	fix splitting of queries when step is larger than split interval (#5006 )	4 years ago
Owen Diehl	1f6453fb81	exposes loki codec (#3758 )	5 years ago
Cyril Tovena	93f2b2d89b	Fixes split interval for metrics queries. (#3459 ) The step needs to be taken into account: - to be removed on each split end. - to not split queries that have a step lower than the split interval. The code is taken from Cortex but adjusted to keep the nanoseconds precision. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com>	5 years ago
Aditya C S	5b8d9b4ed6	Support series query in query-frontend (#2123 )	6 years ago
Owen Diehl	89d80a6fc4	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 `55d41b9519`. * 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
Aditya C S	f9ff2a4b02	Add goimport local-prefixes configuration to .golangci.yml (#1710 )	6 years ago
Owen Diehl	3e075497cd	Feature/per tenant splitby (#1565 ) * split by overrides by tenant * tenant aware metric query splitting * consistent split_queries_by_interval naming * updates readme * fixes rebase conflict * refactor to support limit based splitby intervals in all locations * simplifies limits cachesplitter * fixes comment	6 years ago
Owen Diehl	a1d090e161	Fix/deadlock frontend queue (#1490 ) * buffers splitby interval channels to prevent deadlocks * improves codec merge test requiring merging from multiple intervals * bound concurrency by min(parallelism,jobs) * approximates goroutine test	6 years ago
Cyril Tovena	4153740a12	Loki Query Frontend (#1442 ) * Adds frontend to Loki. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Improves tests. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes sneaky bug in entries sorting. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes the split by interval. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Tweak jsonnet deployments and add a way to lint/fmt. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * lint. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * fix timezone issue. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Improve tests and rollback change in loghttp package. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes a flaky test that might run one more goroutine. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes windows build. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Improve tracing in the split by interval. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Add test stream to proto conversion. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Fixes flappy retry test. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Remove err shadowing in stopQueryFrontend as it was confusing. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Refactor grpc message size in the libsonnet config file. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Don't check auth header for GRPC TransferChunks. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Query frontend (#3) * frontend codec merging optimizations * codec benchmarks * removes unused bounds code in queryrange ordering * [wip] splitby uses channels instead of sub batching intervals * splitBy channel limit test * single allocation for merging entries from a single stream * skip merging loki responses when limit is already hit * removes checks for unlimited queries in queryrange * removes splitByInterval{,.interval} spans * removes interval_batch_size from jsonnet lib * moves benchmark utils to own file * renames markers -> entries * priority queue comments * Removes unused logRequest. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Sets the cache interval to the same split interval. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Missing import libsonnet for the frontend. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> * Frontend should not be a cluster IP. Signed-off-by: Cyril Tovena <cyril.tovena@gmail.com> Co-authored-by: Owen Diehl <ow.diehl@gmail.com>	6 years ago

20 Commits (d24fe3e68be4be16b3654e6aedafaa71ac9d307c)