mirror of
https://github.com/GreptimeTeam/greptimedb.git
synced 2026-11-14 06:46:56 +00:00
discord9
eeca48bc43
I hereby agree to the terms of the [GreptimeDB CLA](https://github.com/GreptimeTeam/.github/blob/main/CLA.md). ## Refer to a related PR or issue link (optional) - Spec: Narthex Unified SeriesFrame refactor, PR A of two. - #9914, #9868, #9937, #9896, #9933, #9936, #9882 and #9886 are all merged, so this is a plain increment on `narthex` with no stacked dependency. ## What's changed and what's your intention? SeriesFrame is the internal transport of the Narthex native PromQL path, and it could only carry raw samples: an evaluated result had to fall back to a plain row table. This makes **one self-describing Arrow format carry both**, so a receiver decodes a frame from the frame alone, without the scanner, region, or planning context its producer had. What is unified is the **physical representation and its codec**, not the runtime data structures. `RawSeriesChunk`, `Series` and `EvaluatedChunk` are unchanged; only the adapter boundaries branch on the sample shape. ### Base - base: `narthex` @ `771347df5d9` - HEAD: `8bc53c48714` - **no stacked dependency**: everything this PR built on has landed This is a plain increment on the current `narthex` tip, with focused follow-up fixes for review and CI. Nothing from the earlier stack is carried: #9914 is merged, so no `MetricName` rename appears here. The codec itself, the adapters, the tests and the two frozen wire fixtures are all that moved, plus the call sites that had to follow the renamed API. ### The format (`layout = v2`) One frame family, three legal combinations. Every other combination is refused on both the writing and the reading side: | content | identity | time | producer | | --- | --- | --- | --- | | `raw` | `key` | `timestamps` | a storage raw-series scan | | `raw` | `explicit` | `timestamps` | a raw input whose labels are already decoded | | `evaluated` | `explicit` | `grid_spans` | a compute node's output | `evaluated` + `key` is refused because a computation may rewrite its labels, so it cannot claim a storage key as its identity. `raw` + `grid_spans` and `evaluated` + `timestamps` are refused because raw samples have no grid and computed values have no timestamps of their own. Column layout (`series_frame_schema` in `src/datatypes/src/series/frame.rs`): ```text read_key: Binary (NOT NULL for key mode; NULLABLE for explicit mode) is_last: Boolean NOT NULL time: Struct<timestamps: List<Int64>> // raw Struct<spans: List<Struct<start: u64, len: u64>>> // evaluated fields: Struct<<field_name>: SamplePayload, ...> name: Utf8 NULLABLE // explicit only, __name__ labels: List<Struct<name, value>> // explicit only, no __name__, no duplicates ``` `read_key` must be populated for every raw row, including raw/explicit, and must be NULL for evaluated rows. `SamplePayload` (float / native histogram / mixed) keeps its existing representation and is not redesigned. Protocol metadata lives under `greptime.series.*`. `SeriesFrameDeclaration::parse` reads the layout, identity, time and ordering declaration. Schemas explicitly declare `contiguous` by default; raw streams may declare `read_key_then_time`, while evaluated streams accept only `contiguous`. Missing ordering, unknown values and unknown reserved keys are refused by declaration parsing, direct decoding (including empty batches), and stream admission. | key | value | | --- | --- | | `layout` | `v2` | | `content` | `raw` / `evaluated` | | `labels` | `key` / `explicit` | | `time` | `timestamps` / `grid_spans` | | `grid` | required for evaluated, forbidden for raw; canonical `GridSpec` JSON | | `label_decode` | required for key, forbidden for explicit | | `ordering` | `contiguous` / `read_key_then_time` | ### Time axis - raw: every row carries its own strictly increasing `i64` millisecond timestamps. - evaluated: the `GridSpec` is declared once in schema metadata (`{start_ms, step_ms, len}`), and a row carries only the **absolute grid positions** as spans. A sparse grid is never expanded and its holes are never filled; `start + len` is checked against the grid with checked arithmetic. - A zero-sample closing chunk is legal for evaluated and **is kept** as one frame row with zero positions and `is_last = true`, because dropping it would drop the finality it exists to carry. A zero-sample chunk that closes nothing is refused, and a raw frame still refuses an empty chunk. ### Series chains and `is_last` `is_last` closes **one chunk chain**, not the end of a PromQL label set. After a native label rewrite, two different input series can produce **the same complete labels**, and each closes itself. The writer allows that and the reader does not treat "same labels" as a re-opened series. A chain that changes identity before closing, and EOF with an open chain, are both errors. ### Key code locations - `src/datatypes/src/series/frame.rs`: shared schema, declaration parser, bounded builder, cursor and raw/evaluated decoder. - `src/common/query/src/series_frame.rs`: typed pack/unpack adapters and chain validators. - `src/datatypes/src/series/time.rs`: canonical `GridSpec` serialization. ### Review and CI fixes - Repaired the two table PK-binding fixtures with real `host` and `rack` Tag columns and matching primary-key metadata. Production `LabelDecodeSpec::validate()` stays strict. The mismatch test retains all six cases, including changed tag order, and checks rejection before polling. - Removed the missing-ordering fallback and added one private ordering validator shared by the declaration parser and direct decoder. Invalid evaluated ordering cannot bypass stream checks. - Compatibility metadata is tested exactly as written, without merging canonical metadata into malformed inputs. The direct-decoder matrix covers raw/evaluated, empty/nonempty, missing, unknown and interleaved ordering, and valid raw versus invalid evaluated `read_key_then_time`. - Independent source and anti-over-design review: **APPROVE** for `771347df5d9..8bc53c48714`. Review approval is separate from the CI status of the new head. ### Limitations - Not reachable from a query yet: `FusedSeriesExec` still answers `PlainRows` by default. The second output format is PR B. - Not verified: Windows; the results below are Linux. - Deliberately **not** done, per the spec: evaluated **input** wiring in `FusedSeriesExec`, Complete Aggregate's SeriesFrame input, Partial/Final physical plans, cross-region exchange, remote NativeTree serialization, NodePath rework, #9892 storage-interface changes, #9844 remote read wiring. - The unreleased v1 frame format is **not** read. In-tree producers, consumers, tests and the two frozen wire fixtures all move to v2 together. ### Does this break API or data compatibility? No published format changes. SST, WAL, manifest and the generic Flight/Protobuf formats are untouched. The only versioned thing here is the unreleased internal SeriesFrame layout, which goes from `v1` to `v2` in one step and is not read in both directions. ### Tests The existing tests include real pack → batch stream → unpack round trips, not only schema snapshots: - raw/key with its decoded PK, raw/explicit with complete labels, both across batches and through Arrow IPC - evaluated dense, sparse (multiple spans, holes not filled), multi-field (a position can hold a value in one field and nothing in another; NULL ≠ NaN), native histogram and mixed payloads (payload kind and values survive, no silent float-ification) - evaluated empty closing chunk kept as one row; multi-batch streams keeping order and finality; two consecutive chains with equal labels each closing independently; EOF with an open chain - malformed frames refused: non-increasing timestamps, overlapping/out-of-order/out-of-bounds/zero length spans, payload length mismatch, span-count overflow, missing grid, malformed `GridSpec`, missing label decode, contradictory metadata, wrong layout version, an evaluated row carrying a storage key, an empty row that closes nothing <!-- CURRENT_VERIFICATION_START --> Verified on Linux at `8bc53c48714`; every command below exited 0: ```text make fmt-check -> clean (rustfmt and repository source checks) cargo nextest run -p datatypes -p common-query -p table --retries 0 -> 690 passed, 0 skipped; includes both reported CI failures cargo nextest run -p mito-codec -p promql -p query --retries 0 -> 2055 passed, 2 skipped cargo nextest run -p mito2 -E 'test(engine::scan_test::test_mito_handle_query_raw_series) | test(engine::scan_test::test_two_phase_native_raw_stream) | test(engine::scan_test::test_two_phase_raw_) | test(engine::scan_test::test_mito_raw_series_)' --retries 0 -> 8 passed, 1733 skipped (focused raw-scan selection) cargo check --workspace --all-targets -> passed cargo clippy -p datatypes -p common-query -p table -p mito-codec -p promql -p query --all-targets -- -D warnings -> passed (scoped crates, default features) ``` License headers, TOML formatting and `git diff --check` also pass. These are local Linux results, not a claim that the new-head GitHub checks are green. Full `make test`, workspace all-features Clippy and Windows execution were not run locally in this round. <!-- CURRENT_VERIFICATION_END --> ## PR Checklist Please convert it to a draft if some of the following conditions are not met. - [x] I have written the necessary rustdoc comments. - [x] I have added the necessary unit tests and integration tests. - [ ] This PR requires documentation updates. - [x] API changes are backward compatible. - [x] Schema or data changes are backward compatible. - [ ] This PR needs to be backported to release branches, and I have added the `backport-<target>` labels (e.g. `backport-v1.3` targets `release/v1.3`).
Metrics, logs, and traces.
One engine, on your infrastructure.
A columnar database for metrics, logs, and traces on object storage. Apache-2.0 licensed core.
User Guide · API Docs · Roadmap 2026 · Slack
stable for production · canary includes pre-releases · nightly is a weekly snapshot of main
Introduction
GreptimeDB is an open-source observability database. Metrics, logs, and traces run on one columnar engine over object storage and share one table model: tags, timestamp, and fields. When signals carry common identifiers such as service, host, or trace ID, you can correlate them in SQL without moving data between databases.
Ingest through OpenTelemetry, Prometheus Remote Write, Loki Push, or Elasticsearch Bulk. Use SQL across observability data and PromQL for metrics. Migrate ingestion one signal at a time without rebuilding your collectors.
One Query Across Signals
OpenTelemetry ingestion writes spans to opentelemetry_traces and log records to
opentelemetry_logs. Both tables carry trace_id, so correlating them is a join:
-- The slowest failed spans in the last hour,
-- with the log lines emitted inside those same traces.
SELECT
t.service_name,
t.span_name,
t.duration_nano / 1000000 AS duration_ms,
l.timestamp AS log_time,
l.severity_text,
l.body
FROM opentelemetry_traces t
JOIN opentelemetry_logs l ON l.trace_id = t.trace_id
WHERE t.timestamp > now() - INTERVAL '1' HOUR
AND t.span_status_code = 'STATUS_CODE_ERROR'
ORDER BY t.duration_nano DESC
LIMIT 20;
Metrics join the same way, on any tag the tables share, such as service,
host, or pod.
Why You Might Use It
- You run Prometheus plus Loki or Elasticsearch and want one backend instead of three
- You have outgrown Prometheus on cardinality or retention and don't want the Thanos/Mimir operational surface
- You are hitting Loki's query performance limits as log volume grows
- You need long retention on object storage without a separate analytics stack
- You want to query telemetry with SQL, not only a domain query language
- You are storing GenAI or agent telemetry (OTel GenAI conventions) alongside infrastructure signals
Learn more in Why GreptimeDB.
What's Supported
| Ingest | OpenTelemetry (OTLP), Prometheus Remote Write, Loki Push, Elasticsearch Bulk, InfluxDB line protocol, gRPC |
| Query | SQL, PromQL, Jaeger-compatible trace queries, MySQL and PostgreSQL wire protocols |
| Storage | S3, GCS, Azure Blob and S3-compatible endpoints as primary storage, with memory and local-disk caches |
| Built in | Retention policies, downsampling, continuous aggregation, explicit table partitioning, and inverted / skipping / fulltext indexes |
Compute and storage are disaggregated: object storage holds the data, while memory and local-disk caches keep recent and frequently queried data close to compute.
Benchmarks
- Agent RCA Bench: LLM agents doing root cause analysis over GreptimeDB versus Prometheus + Loki + Tempo. 40% fewer wrong diagnoses, 48% fewer input tokens, 45% lower cost (write-up)
- GreptimeDB tops JSONBench's billion-record cold run test
- TSBS Benchmark
- More benchmark reports
Compatibility and Migration
Compatibility is per protocol, and query-side coverage is narrower than ingestion.
| Compatible | Not compatible | |
|---|---|---|
| Prometheus | Remote Write ingestion; PromQL queries | Gaps are listed in PromQL compatibility |
| Loki | Push ingestion; dual-write through Grafana Alloy makes the cutover gradual | LogQL and the rest of the Loki query API |
| Elasticsearch | _bulk ingestion in the open-source core; QueryDSL partially, in Enterprise |
Most other Elasticsearch APIs |
Limitations and Edition Boundary
Cluster deployment, object storage, the Flow engine, and every ingestion protocol listed above are in the Apache-2.0 build. Repartitioning, region migration, and index creation are manual operations there.
Read replicas, workload isolation, and automated repartitioning are GreptimeDB Enterprise features, along with enterprise security and governance. The Enterprise overview has the current list, and pricing has the edition comparison.
Architecture
GreptimeDB can run in two modes:
- Standalone — single binary for development and small deployments.
- Distributed — four components, each independently scalable:
- Frontend — protocol entry (OTel, Prometheus, MySQL/PostgreSQL, gRPC, ingestion APIs for Elasticsearch/InfluxDB/Loki) and the distributed query engine. Stateless, scales horizontally.
- Datanode — region engine with WAL, memtable, SST, cache, compaction, and indexes. Persists data to object storage. Elastic.
- Metasrv — metadata, routing, repartitioning, and security. Backed by a pluggable KV layer (etcd or RDS).
- Flownode (optional) — continuous flow computation (streaming and materialized views).
For deeper coverage, see the architecture doc or DeepWiki.
Try GreptimeDB
For AI agents — paste this prompt into your agent:
Read https://docs.greptime.com/SKILL.md and follow the instructions
to deploy, configure, ingest, and query GreptimeDB.
docker run -p 127.0.0.1:4000-4003:4000-4003 \
-v "$(pwd)/greptimedb_data:/greptimedb_data" \
--name greptime --rm \
greptime/greptimedb:latest standalone start \
--http-addr 0.0.0.0:4000 \
--grpc-bind-addr 0.0.0.0:4001 \
--mysql-addr 0.0.0.0:4002 \
--postgres-addr 0.0.0.0:4003
Dashboard: http://localhost:4000/dashboard
Read more in the full Install Guide.
Troubleshooting:
- Cannot connect to the database? Ensure that ports
4000,4001,4002, and4003are not blocked by a firewall or used by other services. - Failed to start? Check the container logs with
docker logs greptimefor further details.
Getting Started
Build From Source
Prerequisites:
- Rust toolchain — stable, pinned by
rust-toolchain.toml - Protobuf compiler (>= 3.15)
- C/C++ building essentials:
gcc/g++/autoconfand the glibc dev package (libc6-devon Ubuntu,glibc-develon Fedora) - Python toolchain (optional, only for some test scripts)
Build and run:
make # build greptime binary
cargo run -- standalone start # start in standalone mode
Common dev commands:
make fmt # format Rust code
make clippy # lint (fails on warnings)
make test # unit + integration tests (uses cargo-nextest)
make sqlness-test # SQL regression tests
See the Contribution Guidelines for the full developer workflow.
Tools & Extensions
- Kubernetes: GreptimeDB Operator
- Helm Charts: Greptime Helm Charts
- Dashboard: Web UI
- gRPC Ingester: Go, Java, C++, Erlang, Rust, .NET, TypeScript
- Grafana Data Source: GreptimeDB Grafana data source plugin
- Grafana Dashboard: Official Dashboard for monitoring
Project Status
GreptimeDB is generally available, with stable APIs and regular releases. It runs in production at scale — OceanBase Cloud operates 80+ GreptimeDB clusters managing 300 TB of logs, cutting log storage cost by 60%+ after migrating from Grafana Loki. See more in case studies.
Release lines and support windows are in the version reference. For where the project is going, read the v1.0 highlights and the 2026 roadmap.
Community
We invite you to engage and contribute!
If GreptimeDB is useful to you, please star the repo.
License
GreptimeDB is an open-core project. Its core is licensed under the Apache License 2.0.
A small set of peripheral, enterprise-only features are gated behind the
enterprise Cargo feature (not built by default) and are governed by the
separate GreptimeDB Enterprise License. Source files under
that license carry an explicit Enterprise License header.
Commercial Support
Scaling observability on your infrastructure? GreptimeDB Enterprise adds the operational, security, and support layer for production deployments. Contact us for details.
Contributing
- Read our Contribution Guidelines.
- Explore Internal Concepts and DeepWiki.
- Pick up a good first issue and join the #contributors Slack channel.
Acknowledgement
Special thanks to all contributors! See AUTHOR.md.
- Uses Apache Arrow™ (memory model)
- Apache Parquet™ (file storage)
- Apache DataFusion™ (query engine)
- Apache OpenDAL™ (data access abstraction)
All trademarks, logos, and brand names referenced in this README and in the Overview diagram are the property of their respective owners. Their use is for identification purposes only and does not imply endorsement or affiliation.