129 Commits (5f3300a0848380bb901c44148ffdc43703fc4717)

Author	SHA1	Message	Date
Christian Haudum	b316d52612	chore(engine): Use shard information from query frontend in the new query engine (#17792) ... **What this PR does / why we need it**: This PR introduces the sharding to the new query engine to be able to test the engine using the exsiting Loki architecture with query frontend, query scheduler, and queriers. **Note, this is only an interim solution used for validating and testing.** By design, the first phase of the query engine implementation does only local execution of queries without sharding or time-based splitting. However, for testing the results of this first phase, we can utilise dual-ingestion (writing both chunks/tsdb and data objects in a way that sharding and splitting is performed in the query frontend using the existing middlewares. On the queriers themselves, the sub-queries are parsed and planned using the new logical and physical planner. If the query can successfully be planned, it will be executed by the new engine, otherwise it falls back to the old engine. **Shard size considerations** While performing time-splitting in the frontend works for data objects as well, sharding by information from TSDB is not directly mappable to data objects. The default target shard size in TSDB is 600MB (decompressed), whereas target size of data objects is 1GB compressed or roughly 10-15GB uncompressed. However, individual logs sections of a data object have a target size of 128MB, which is roughly 0.9-1.2GB. That is 1.5-2x larger than the TSDB target shard size. So when using the sharding calculation from TSDB, it would over-shard for data object sections, which is likely acceptable for testing and good enough for proving that local execution with the new engine works. **How does sharding with data objects in this PR work?** The query frontend passes down the calculated shards as part of the query parameters of the serialised sub-request. The logical planner on the querier stores the Shard annotation on the `MAKETABLE` alongside the stream selector. This is then used by the physical planner to filter out only the relevant sections of the resolved data objects from the metastore lookup. During exeuction, only readers for the relevant sections are initialised when performing the `DataObjScan`. Signed-off-by: Christian Haudum <christian.haudum@gmail.com> Co-authored-by: Ashwanth <iamashwanth@gmail.com>	7 months ago
Robert Fratto	06cf52757b	chore(dataobj): Add columnar reading APIs to logs and streams sections (#17976) ... This PR introduces columnar reading APIs to both the logs and streams sections. As part of this change, both sections now expose the columns stored in the section, which are then used to define predicates on the readers. The columnar reading APIs emit a sequence of `arrow.Record`s, where each record has no more than the batch size passed to the `Reader.Read` method. Unlike the original row-based readers, the columnar readers: * Allow reading a subset of columns * Have column-based predicates that map literally to the new query engine To be able to represent a column's value for use in predicates, we import [github.com/apache/arrow-go/v18/arrow/scalar](https://pkg.go.dev/github.com/apache/arrow-go/v18@v18.3.0/arrow/scalar). Section Reader implementations perform mapping to and from the internal `dataset.Value`. The implementation of `logs.Reader` is almost an identical copy of the `streams.Reader` implementation. I've opted for duplicating the implementation in the short term, since there's no obvious way for how to deduplicate it in a way keeps the package structure easy to understand. We can change this in the future if a clean and simple approach presents itself. This PR does not yet update `DataObjScan` to make use of these new APIs; that's being left out of scope for another PR to handle.	7 months ago
Robert Fratto	c5ac922e55	chore(dataobj): Use pooled buffers when decoding pages (#17955)	8 months ago
Charles Korn	545fb1597c	refactor(stringlabels): Support stringlabels in `logql/log` package (#17838) ... This is another step towards supporting Prometheus' stringlabels implementation in Loki. It adds support in the `logql/log` package. Part of https://github.com/grafana/loki/issues/17122 The tests should now compile and pass with `-tags stringlabels`. --- Co-authored-by: Karsten Jeschkies <karsten.jeschkies@grafana.com>	8 months ago
Christian Haudum	48f1d0ebc0	chore(dataobj): Decompress zstd compressed pages upfront (#17841) ... This PR changes the decompression of data object pages, so that they are decompressed fully upfront into a bytes buffer, instead of passing a streaming zstd decompressor (reader) to the page decoder. This this allows to release the zstd decompressor to the pool earlier.Signed-off-by: Christian Haudum <christian.haudum@gmail.com>	8 months ago
Oleg Zaytsev	c8a15f451c	refactor: Use OTel tracing library (#17859) ... Signed-off-by: Oleg Zaytsev <mail@olegzaytsev.com>	8 months ago
benclive	e6fe98c7c3	chore: Emit metastore event to kafka after object build (#17842)	8 months ago
benclive	02a6b81c66	chore: Estimate builder size when using multiple sections (#17840)	8 months ago
Trevor Whitney	3d2286a840	chore: implement multi-variant extractor that can return multiple samples (#17149)	8 months ago
Jackson Coelho	8f58df6e47	chore: rename dataobj consumer metrics (#17830)	8 months ago
Jackson Coelho	8001beef13	chore: add metrics (#17823)	8 months ago
Robert Fratto	36bcb5aa2d	chore(dataobj): fix regression in allocations of row readers (#17784)	8 months ago
Robert Fratto	a237444f0d	refactor(dataobj): invert dependency between dataobj and sections (#17762) ... Originally, the dataobj package was a higher-level API around sections. This design caused it to become a bottleneck: * Implementing any new public behaviour for a section required bubbling it up to the dataobj API for it to be exposed, making it tedious to add new sections or update existing ones. * The `dataobj.Builder` pattern was focused on constructing dataobjs for storing log data, which will cause friction as we build objects around other use cases. This PR builds on top of the foundation laid out by #17704 and #17708, fully inverting the dependency between dataobj and sections: * The `dataobj` package has no knowledge of what sections exist, and can now be used for writing and reading generic sections. Section packages now create higher-level APIs around the abstractions provided by `dataobj`. * Section packages are now public, and callers interact directly with these packages for writing and reading section-specific data. * All logic for a section (encoding, decoding, buffering, reading) is now fully self-contained inside the section package. Previously, the implementation of each section was spread across three packages (`pkg/dataobj/internal/encoding`, `pkg/dataobj/internal/sections/SECTION`, `pkg/dataobj`). * Cutting a section is now a decision made by the caller rather than the section implementation. Previously, the logs section builder would create multiple sections. For the most part, this change is a no-op, with two exceptions: 1. Section cutting is now performed by the caller; however, this shouldn't result in any issues. 2. Removing the high-level `dataobj.Stream` and `dataobj.Record` types will temporarily reduce the allocation gains from #16988. I will address this after this PR is merged.	8 months ago
Robert Fratto	7b0d6d7144	chore(dataobj): Migrate to dynamic section types (#17708)	8 months ago
Robert Fratto	3fbebd3bc4	chore(dataobj): migrate to section-relative offsets (#17704)	8 months ago
Ashwanth	3ad1a64bc1	chore(dataobj): fixes a memory aliasing bug in dataobj reader (#17537)	8 months ago
Trevor Whitney	577d501839	refactor: extractors can return multiple samples (#17064) ... Signed-off-by: Trevor Whitney <trevorjwhitney@gmail.com>	9 months ago
Robert Fratto	4841b2abe2	chore(engine): provide basic implementation of DataObjScan (#17568) ... Co-authored-by: Ashwanth Goli <iamashwanth@gmail.com>	9 months ago
Ashwanth	009600cb51	chore(dataobj): order predicates based on row selectivity and cost of reading (#17376)	9 months ago
Ashwanth	5961923836	chore(dataobj): apply predicates sequentially (#17283)	9 months ago
Ashwanth	42e2407137	chore(dataobj): backfill sparse rows using EncodeN (#16485)	9 months ago
aarogoss	ab65ac5e8e	feat: refactors kafka/client to improve metric collection (#17308)	9 months ago
Ashwanth	7dd4ab4b09	chore(dataobj): correct stats calculation for prepredicate bytes read (#17279)	9 months ago
Karsten Jeschkies	a864f8b986	refactor(stringlabels): Support stringlabels in log.StreamPipeline (#17216)	9 months ago
benclive	9e68fb01f4	feat(dataobj): Add query stats collection to the dataobj readers (#17128)	9 months ago
benclive	e5784d7a82	perf: Reuse Values in dataobj.Reader to reduce allocs (#16988) ... Co-authored-by: Robert Fratto <robertfratto@gmail.com>	9 months ago
Robert Fratto	bd4331363b	chore: align benchmark results between chunks and dataobjs (#17127)	9 months ago
benclive	8e421186df	chore: Collect basic stats from dataobj queries (#17111)	9 months ago
aarogoss	e962398462	feat: Optionally enable kafka histograms for read/write latency (#17089)	9 months ago
Christian Haudum	e297c51d20	chore(metastore): Use constants for label names (#17048) ... This PR contains improvements for: 1. Make label names contants 2. Use `strconv.FormatInt` for encoding timestamp into label value (equivalent to `strconv.ParseInt` when decoding) 3. Use `labels.New` to construct labels, rather then string formatting Signed-off-by: Christian Haudum <christian.haudum@gmail.com>	10 months ago
Christian Haudum	ba61aa88de	chore(engine): Integrate metastore into physical query planning (#17016) ... This PR integrates the data object lookup using the metastore into the query planning. To achieve that, I extended the `Metastore` interface with an additional function that not only returns the data object paths, but also the stream IDs of streams that match the given label matchers. ```go type Metastore interface { ... // StreamsIDs returns object store paths and stream IDs for all matching objects for the given matchers between [start,end] StreamIDs(ctx context.Context, start, end time.Time, matchers ...*labels.Matcher) ([]string, [][]int64, error) ... } ``` Signed-off-by: Christian Haudum <christian.haudum@gmail.com>	10 months ago
Ashwanth	42eed6ded5	chore(dataobj): introduce InPredicate to replace long chain of OR operators (#16919)	10 months ago
Christian Haudum	0a3230f457	chore(ast): Update WalkFn signature to return bool (#16989) ... The boolean return value indicates whether walking/traversing should be continued with child nodes. This PR does not change any existing traversal behaviour. Signed-off-by: Christian Haudum <christian.haudum@gmail.com>	10 months ago
Robert Fratto	9be95a3e8c	chore(dataobj): reduce memory overhead of merging logs stripes (#16833)	10 months ago
Ashwanth	c51fa39684	chore(dataobj): pass down line filters to logs reader (#16802)	10 months ago
benclive	4befd22dd9	feat(dataobj-explorer): Add stream distribution info to dataobj explorer UI (#16525)	10 months ago
aarogoss	a57a80ea32	feat(metastore): Implement Labels() and Values() on ObjectMetastore (#16734)	10 months ago
Christian Haudum	9691f3e5dc	chore: Move `pkg/dataobj/planner` to `pkg/engine/planner` (#16815) ... The package pkg/dataobj/planner/logical was initially created within the package pkg/dataobj because of the dependency on the internal package pkg/dataobj/internal/metadatamd. However, the package should be located inside a separate package without the dependency on the dataobj package, since the engine should be storage agnostic. --- Signed-off-by: Christian Haudum <christian.haudum@gmail.com>	10 months ago
benclive	979f7614bc	perf: Optimization pass for reading from dataobjs (#16747)	10 months ago
Christian Haudum	f2257f5be0	chore: Implement the logical plan formatter using the generic tree printer (#16761) ... This PR implements the logical plan specific tree formatter using the generic tree printer introduced with https://github.com/grafana/loki/pull/16716 This solves the problem that expressions, which are properties of nodes, were treated equally as the nodes' child node. --- Signed-off-by: Christian Haudum <christian.haudum@gmail.com>	10 months ago
Christian Haudum	9bdd5d15bf	chore: Add tree printer for physical plan (#16716) ... This PR adds a generic tree printer (similar to the Unix utility `tree`). Additionally it provides an implementation to convert the DAG of the physical plan into the generic tree used by the tree printer. This allows to print plans like this (note, these examples do not make sense, they are only for demonstrating the visual output of the printer). Simple DAG with single child node: ``` Limit #limit offset=0 limit=0 └── Filter #filter predicates=() └── SortMerge #merge column=<nil> order=UNDEFINED ├── DataObjScan #scan1 location= stream_ids=() projections=() predicates=() direction=0 limit=0 └── DataObjScan #scan2 location= stream_ids=() projections=() predicates=() direction=0 limit=0 ``` DAG with multple root nodes: ``` Limit #limit1 offset=0 limit=0 └── DataObjScan #scan1 location= stream_ids=() projections=() predicates=() direction=0 limit=0 Limit #limit2 offset=0 limit=0 └── DataObjScan #scan2 location= stream_ids=() projections=() predicates=() direction=0 limit=0 ``` DAG with parent node that share the same child node: ``` Limit #limit offset=0 limit=0 ├── Limit #filter1 offset=0 limit=0 │ └── DataObjScan #scan location= stream_ids=() projections=() predicates=() direction=0 limit=0 └── Limit #filter2 offset=0 limit=0 └── DataObjScan #scan location= stream_ids=() projections=() predicates=() direction=0 limit=0 ``` --- Signed-off-by: Christian Haudum <christian.haudum@gmail.com>	10 months ago
Christian Haudum	5a76a04684	chore: Add basic building blocks of a physical plan (#16710) ... This PR adds the basic components of a physical plan. For the first iteration of the new execution engine, we need to support nodes to represent limited and filter queries (with only label/line filters). These nodes are: * DataObjScan * SortMerge * Projection * Filter * Limit The physical plan is implemented as a [directed acyclic graph](https://en.wikipedia.org/wiki/Directed_acyclic_graph) (DAG). It has convenience functions for adding nodes (vertices) and edges, getting parents and children of a node, as well as for traversing the graph using the [depth-first-search](https://en.wikipedia.org/wiki/Depth-first_search) algorithm. Additionally to the node types, the PR also contains the basic types of expressions, which can be properties of the nodes. --- Signed-off-by: Christian Haudum <christian.haudum@gmail.com>	10 months ago
Robert Fratto	cd11367148	chore(dataobj): Wire dataset.Reader into dataobj.LogsReader and dataobj.StreamsReader (#16701) ... Co-authored-by: Ashwanth Goli <iamashwanth@gmail.com>	10 months ago
Robert Fratto	8c4eb428c3	chore(dataobj): Download pages in 16MB batches (#16689)	10 months ago
Robert Fratto	abf3772366	perf(dataobj): improve parallelized read spead of Zstd (#16632)	10 months ago
benclive	91bab5e7f2	chore(metastore): Fix stream read condition (#16631)	11 months ago
Owen Diehl	89f0ed795f	feat(logical planning): Planner playground (#16403)	11 months ago
benclive	475d25f459	feat(metastore): Add initial interface for Metastore (#16604)	11 months ago
Paul Rogers	27431b7e7e	chore: Linting update for new golangci (#16572)	11 months ago
Robert Fratto	7f0dc01e21	chore(dataobj): introduce more efficient dataset reading (#16429) ... This commit introduces a new API, dataset.Reader, which scans through values in datasets more efficiently than dataset.Iter. As a result, dataset.Iter has been removed in favour of the new Reader. dataset.Reader is more efficient than dataset.Iter because: - It avoids using iter.Pull; the number of values flowing through coroutines added a significant amount of CPU overhead. - It supports efficient uses of predicates, where non-predicate columns are only loaded after finding matching rows from predicate columns; this avoids loading unnecessary pages. - It adds initial support for bulk page downloading, so that we need fewer roundtrips to object storage when iterating over a dataset. (This is not fully implemented yet; see below). - Early performance benchmarks see a ~3.4x speed improvement over reading 7GB of uncompressed logs, going from 192 MB/s (with dataset.Iter) to 652 MB/s (with dataset.Reader). This commit is truly massive, and to avoid it getting any larger, I'm leaving some follow up work to additional commits: - Updating dataobj.LogsReader and dataobj.StreamsReader to use the new API. - Adding support for downloading pages in batches; dataset.Reader will request multiple pages at once, but our decoder still unwraps that into several GET requests to object storage. Signed-off-by: Robert Fratto <robertfratto@gmail.com> Co-authored-by: Cyril Tovena <cyril.tovena@gmail.com>	11 months ago