* fix(server): describe EXPLAIN statements so bind parameters work
`do_describe_inner` only planned `Insert`/`Query`/`Delete`, so
`EXPLAIN` and `EXPLAIN ANALYZE` fell through to the non-plan branch
and had no parameter-type inference. At Bind time the Postgres
handler then reported `unsupported_parameter_type` even though the
inner query would have worked on its own.
Recurse one level into `Statement::Explain` so that an EXPLAIN
wrapping a plannable statement goes through the same describe path.
Adds a tokio-postgres integration test that exercises
`EXPLAIN`/`EXPLAIN ANALYZE` over the extended query protocol.
Fixes#8029
Signed-off-by: BootstrapperSBL <yvanwww@gmail.com>
* refactor(server): extract plannable-inner check into closure
Reduce duplication between the direct match and the EXPLAIN inner match
by factoring out is_inner_plannable. Behaviour unchanged.
Signed-off-by: BootstrapperSBL <yvanwww@gmail.com>
---------
Signed-off-by: BootstrapperSBL <yvanwww@gmail.com>
* feat: add support for decimal parameter type, remove string replacement fallback
* chore: format
* fix: add support for using unsigned bigint in postgres
* chore: format toml
* refactor: cleanup duplicated code
* fix: rescale decimal
* feat: support alter from primary_key to flat
Signed-off-by: evenyag <realevenyag@gmail.com>
* chore: alter flat to primary_key
Signed-off-by: evenyag <realevenyag@gmail.com>
* feat: change default_experimental_flat_format to true
Signed-off-by: evenyag <realevenyag@gmail.com>
* feat: compute channel size from splitted batch size
Signed-off-by: evenyag <realevenyag@gmail.com>
* test: add tests for split and channel size
Signed-off-by: evenyag <realevenyag@gmail.com>
* fix: always set sst_format from manifest on region open
sanitize_region_options did not set options.sst_format when the
default (PrimaryKey) matched the manifest value, leaving it as None
after reopen. This caused the alter format change to appear lost.
Signed-off-by: evenyag <realevenyag@gmail.com>
* test: fix tests
Signed-off-by: evenyag <realevenyag@gmail.com>
* test: show create table after alteration
Signed-off-by: evenyag <realevenyag@gmail.com>
* refactor!: rename default_experimental_flat_format to default_flat_format
The flat format is no longer experimental. Remove "experimental" from
the config field name, doc comments, and all references.
Signed-off-by: evenyag <realevenyag@gmail.com>
* chore: fix clippy
Signed-off-by: evenyag <realevenyag@gmail.com>
---------
Signed-off-by: evenyag <realevenyag@gmail.com>
* feat: metric batch 2s PoC
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* chore: max_concurrent_flushes
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* chore: work channel size
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
* feat(servers): add metrics and logs for pending rows batch flush
Add the `FLUSH_ELAPSED` histogram metric to track the duration of pending
rows batch flushes in the Prometheus store protocol handler. This provides
better observability into the performance and latency of the batcher.
Also update telemetry by:
- Recording elapsed time for both successful and failed flush operations.
- Adding an informational log upon successful flush including row count and duration.
- Including elapsed time in error logs when a flush fails.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat(servers): implement columnar batching for pending rows
Refactor PendingRowsBatcher to use columnar batching for the metrics
store. Incoming RowInsertRequests are now converted to RecordBatches,
partitioned, and flushed via BulkInsert requests to datanodes.
- Enhance MultiDimPartitionRule to handle scalar boolean predicates.
- Add metrics for tracking flush failures and dropped rows.
- Update dependencies to support columnar batching in servers.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat(servers): add backpressure for pending rows
Implement backpressure in PendingRowsBatcher by limiting in-flight
requests with a semaphore and making the submission wait for the flush
result. This ensures Prometheus write requests are throttled and only
return once the data has been successfully flushed to datanodes.
- Add max_inflight_requests to PromStoreOptions.
- Use oneshot channels to notify submitters of flush completion.
- Limit concurrent requests using a new inflight_semaphore.
- Update PendingRowsBatcher::submit to wait for the flush outcome.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat: add stage-level metrics for bulk ingestion
Introduce histograms to track the elapsed time of various stages in the
metric engine bulk insert path and the server's pending rows batcher.
This provides better observability into the performance bottlenecks
of the ingestion pipeline.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* - `src/metric-engine/src/engine/bulk_insert.rs`: Removed the fallback mechanism that converted record batches to rows when bulk inserts were unsupported, along with related helper functions and unused imports.
- `src/operator/src/insert.rs`: Removed an unused import (`common_time::TimeToLive::Instant`).
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat(servers): columnar Prom remote write
Optimize the Prometheus remote write path by allowing direct conversion
from decoded Prometheus samples to Arrow RecordBatches. This bypasses
intermediate row-based representations when `PendingRowsBatcher` is
active and no pipeline is used, improving ingestion efficiency.
- Implement `as_record_batch_groups` in `TablesBuilder` and `PromWriteRequest`.
- Add `submit_prom_record_batch_groups` to `PendingRowsBatcher`.
- Introduce `DecodedPromWriteRequest` in `prom_store`.
- Implement row-to-RecordBatch conversion logic in `prom_row_builder`.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* Revert "feat(servers): columnar Prom remote write"
This reverts commit efbb63c12a3e7fcec03858ea0351efd94fec8242.
* refactor(servers): improve row to RecordBatch conversion
- Use `snafu::ensure` for row validation in `rows_to_record_batch`.
- Add explicit type hint for `MutableVector` to improve clarity.
- Reorganize and clean up imports in `pending_rows_batcher.rs`.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* perf(servers): use arrow builders for row conversion
This commit optimizes the conversion from `api::v1::Rows` to `RecordBatch`
by using Arrow builders directly. This avoids the overhead of
`MutableVector` and `common_recordbatch`, leading to better performance
in the `pending_rows_batcher`.
Additionally, the `#[allow(dead_code)]` attribute is removed from
`modify_batch_sparse` in the metric engine as it is now utilized.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* perf(metric-engine): optimize batch modification
Optimize `modify_batch_sparse` by reusing buffers, using Arrow
builders, and employing fast-path encoding methods. This reduces
allocations and avoids redundant downcasting and serializer overhead.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/metric-engine-support-bulk:
**Add Environment Variable for Batch Sync Control**
- `pending_rows_batcher.rs`: Introduced an environment variable `PENDING_ROWS_BATCH_SYNC` to control the synchronization behavior of batch processing. If set to true, the function will wait for the flush result; otherwise, it will return immediatel
with the total rows count.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* wip
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* chore: update and fix clippy
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* fix: failing test
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* picking-pending-rows-batcher:
### Commit Message
Remove Unused Code and Simplify Error Handling
- **`src/error.rs`**: Removed the `BatcherQueueFull` error variant and its associated logic, simplifying the error handling by removing unused code.
- **`src/http/prom_store.rs`**: Eliminated the `try_decompress` function, streamlining the decompression logic by directly using `snappy_decompress` in `decode_remote_read_request`.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* chore: parse PENDING_ROWS_BATCH_SYNC once
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* chore: revert unrelated changes
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* **Refactor Prometheus Write Handling**
- **`prom_store.rs`**: Introduced `pre_write` method in `PromStoreProtocolHandler` to handle pre-write checks for Prometheus remote write requests. Updated `write` method to utilize `pre_write`.
- **`server.rs`**: Modified `PendingRowsBatcher` initialization to conditionally create a batcher based on `with_metric_engine` flag.
- **`http/prom_store.rs`**: Integrated `pre_write` checks before submitting requests to `PendingRowsBatcher`.
- **`query_handler.rs`**: Added `pre_write` method to `PromStoreProtocolHandler` trait for pre-write operations.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* picking-pending-rows-batcher:
- **Fix Label Typo**: Corrected a typo in the label value from `"flush_wn ite_region"` to `"flush_write_region"` in `pending_rows_batcher.rs`.
- **Refactor Array Building Logic**: Introduced a macro `build_array!` to streamline the construction of `ArrayRef` for different data types, reducing code duplication in `pending_rows_batcher.rs`.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* format toml
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* picking-pending-rows-batcher:
### Update PromStore and PendingRowsBatcher Configuration
- **`prom_store.rs`**: Set `pending_rows_flush_interval` to `Duration::ZERO` to disable automatic flushing.
- **`pending_rows_batcher.rs`**: Enhance validation to disable the batcher when `flush_interval` is zero or configuration values like `max_batch_rows`, `max_concurrent_flushes`, `worker_channel_capacity`, or `max_inflight_requests` are zero, preventing potential panics or deadlocks.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* picking-pending-rows-batcher:
### Update `pending_rows_flush_interval` to Zero
- **Files Modified**:
- `src/frontend/src/service_config/prom_store.rs`
- `tests-integration/tests/http.rs`
- **Key Changes**:
- Updated `pending_rows_flush_interval` from `Duration::from_secs(2)` to `Duration::ZERO` in `prom_store.rs`.
- Changed `pending_rows_flush_interval` configuration from `"2s"` to `"0s"` in `http.rs`.
These changes set the flush interval to zero, potentially affecting how frequently pending rows are flushed.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* picking-pending-rows-batcher:
**Add Worker Management Enhancements**
- **`metrics.rs`**: Introduced `PENDING_WORKERS` gauge to track active pending rows batch workers.
- **`pending_rows_batcher.rs`**:
- Added worker idle timeout logic with `WORKER_IDLE_TIMEOUT_MULTIPLIER`.
- Implemented worker management functions: `spawn_worker`, `remove_worker_if_same_channel`, and `should_close_worker_on_idle_timeout`.
- Enhanced worker lifecycle management to handle idle workers and ensure proper cleanup.
- **Tests**: Added unit tests for worker removal and idle timeout logic.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* fix: clippy
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
---------
Signed-off-by: jeremyhi <fengjiachun@gmail.com>
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
Co-authored-by: jeremyhi <fengjiachun@gmail.com>
* feat: use arrow-pg for encode_row
* refactor: remove bytea and datetime module
* feat: port more encodings to arrow-pg
* feat: implement intervalstyle
* chore: format
* chore: remove error that is no longer used
* chore: use released arrow-pg
* Apply suggestions from code review
Co-authored-by: LFC <990479+MichaelScofield@users.noreply.github.com>
---------
Co-authored-by: LFC <990479+MichaelScofield@users.noreply.github.com>
* feat: impl vector index scan in storage
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* feat: fallback to read remote blob when blob not found
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* chore: refactor encoding and decoding and apply suggestions
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: license
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* test: add apply_with_k tests
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* chore: apply suggestions
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: forgot to align nulls when the vector column is not in the batch
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* test: add test for vector column is not in a batch while buiilding
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
---------
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* feat: impl vector index building
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* feat: supports flat format
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* ci: add vector_index feature to test
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* chore: apply suggestions
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* chore: apply suggestions from copilot
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
---------
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* feat/allow-one-to-many-pipeline:
### Enhance Pipeline Processing for One-to-Many Transformations
- **Support One-to-Many Transformations**:
- Updated `processor.rs`, `etl.rs`, `vrl_processor.rs`, and `greptime.rs` to handle one-to-many transformations by allowing VRL processors to return arrays, expanding each element into separate rows.
- Introduced `transform_array_elements` and `values_to_rows` functions to facilitate this transformation.
- **Error Handling Enhancements**:
- Added new error types in `error.rs` to handle cases where array elements are not objects and for transformation failures.
- **Testing Enhancements**:
- Added tests in `pipeline.rs` to verify one-to-many transformations, single object processing, and error handling for non-object array elements.
- **Context Management**:
- Modified `ctx_req.rs` to clone `ContextOpt` when adding rows, ensuring correct context management during transformations.
- **Server Pipeline Adjustments**:
- Updated `pipeline.rs` in `servers` to handle transformed outputs with one-to-many row expansions, ensuring correct row padding and request formation.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
Add one-to-many VRL pipeline test in `http.rs`
- Introduced `test_pipeline_one_to_many_vrl` to verify VRL processor's ability to expand a single input row into multiple output rows.
- Updated `http_tests!` macro to include the new test.
- Implemented test scenarios for single and multiple input rows, ensuring correct data transformation and row count validation.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
### Add Tests for VRL Pipeline Transformations
- **File:** `src/pipeline/src/etl.rs`
- Added tests for one-to-many VRL pipeline expansion to ensure multiple output rows from a single input.
- Introduced tests to verify backward compatibility for single object output.
- Implemented tests to confirm zero rows are produced from empty arrays.
- Added validation tests to ensure array elements must be objects.
- Developed tests for one-to-many transformations with table suffix hints from VRL.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
### Enhance Pipeline Transformation with Per-Row Table Suffixes
- **`src/pipeline/src/etl.rs`**: Updated `TransformedOutput` to include per-row table suffixes, allowing for more flexible routing of transformed data. Modified `PipelineExecOutput` and related methods to
handle the new structure.
- **`src/pipeline/src/etl/transform/transformer/greptime.rs`**: Enhanced `values_to_rows` to support per-row table suffix extraction and application.
- **`src/pipeline/tests/common.rs`** and **`src/pipeline/tests/pipeline.rs`**: Adjusted tests to validate the new per-row table suffix functionality, ensuring backward compatibility and correct behavior in
one-to-many transformations.
- **`src/servers/src/pipeline.rs`**: Modified `run_custom_pipeline` to process transformed outputs with per-row table suffixes, grouping rows by `(opt, table_name)` for insertion.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
### Update VRL Processor Type Checks
- **File:** `vrl_processor.rs`
- **Changes:** Updated type checking logic to use `contains_object()` and `contains_array()` methods instead of `is_object()` and `is_array()`. This change ensures
compatibility with VRL type inference that may return multiple possible types.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
- **Enhance Error Handling**: Added new error types `ArrayElementMustBeObjectSnafu` and `TransformArrayElementSnafu` to improve error handling in `etl.rs` and `greptime.rs`.
- **Refactor Error Usage**: Moved error usage declarations in `transform_array_elements` and `values_to_rows` functions to the top of the file for better organization in `etl.rs` and `greptime.rs`.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
### Update `greptime.rs` to Enhance Error Handling
- **Error Handling**: Modified the `values_to_rows` function to handle invalid array elements based on the `skip_error` parameter. If `skip_error` is true, invalid elements are skipped; otherwise, an error is returned.
- **Testing**: Added unit tests in `greptime.rs` to verify the behavior of `values_to_rows` with different `skip_error` settings, ensuring correct processing of valid objects and appropriate error handling for invalid elements.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
### Commit Summary
- **Enhance `TransformedOutput` Structure**: Refactored `TransformedOutput` to use a `HashMap` for grouping rows by `ContextOpt`, allowing for per-row configuration options. Updated methods in `PipelineExecOutput` to support the new structure (`src/pipeline/src/etl.rs`).
- **Add New Transformation Method**: Introduced `transform_array_elements_to_hashmap` to handle array inputs with per-row `ContextOpt` in `HashMap` format (`src/pipeline/src/etl.rs`).
- **Update Pipeline Execution**: Modified `run_custom_pipeline` to process `TransformedOutput` using the new `HashMap` structure, ensuring rows are grouped by `ContextOpt` and table name (`src/servers/src/pipeline.rs`).
- **Add Tests for New Structure**: Implemented tests to verify the functionality of the new `HashMap` structure in `TransformedOutput`, including scenarios for one-to-many mapping, single object input, and empty arrays (`src/pipeline/src/etl.rs`).
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
### Refactor `values_to_rows` to Return `HashMap` Grouped by `ContextOpt`
- **`etl.rs`**:
- Updated `values_to_rows` to return a `HashMap` grouped by `ContextOpt` instead of a vector.
- Adjusted logic to handle single object and array inputs, ensuring rows are grouped by their `ContextOpt`.
- Modified functions to extract rows from default `ContextOpt` and apply table suffixes accordingly.
- **`greptime.rs`**:
- Enhanced `values_to_rows` to handle errors gracefully with `skip_error` logic.
- Added logic to group rows by `ContextOpt` for array inputs.
- **Tests**:
- Updated existing tests to validate the new `HashMap` return structure.
- Added a new test to verify correct grouping of rows by per-element `ContextOpt`.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
### Refactor and Enhance Error Handling in ETL Pipeline
- **Refactored Functionality**:
- Replaced `transform_array_elements` with `transform_array_elements_by_ctx` in `etl.rs` to streamline transformation logic and improve error handling.
- Updated `values_to_rows` in `greptime.rs` to use `or_default` for cleaner code.
- **Enhanced Error Handling**:
- Introduced `unwrap_or_continue_if_err` macro in `etl.rs` to allow skipping errors based on pipeline context, improving robustness in data processing.
These changes enhance the maintainability and error resilience of the ETL pipeline.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/allow-one-to-many-pipeline:
### Update `Row` Handling in ETL Pipeline
- **Refactor `Row` Type**: Introduced `RowWithTableSuffix` type alias to simplify handling of rows with optional table suffixes across the ETL pipeline.
- **Modify Function Signatures**: Updated function signatures in `etl.rs` and `greptime.rs` to use `RowWithTableSuffix` for better clarity and consistency.
- **Enhance Test Coverage**: Adjusted test logic in `greptime.rs` to align with the new `RowWithTableSuffix` type, ensuring correct grouping and processing of rows by TTL.
Files affected: `etl.rs`, `greptime.rs`.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
---------
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/change-tsid-gen:
perf(metric-engine): replace mur3 with fxhash for faster TSID generation
- Switches from mur3::Hasher128 to fxhash::FxHasher for TSID hashing
- Pre-computes label-name hash when no nulls are present, avoiding redundant work
- Adds fast-path for rows without nulls; falls back to slow path otherwise
- Updates Cargo.toml and lockfile to reflect dependency change
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/change-tsid-gen:
fix: only check primary-key labels for null when re-using cached hash
- Rename has_null() → has_null_labels() and restrict the check to the
primary-key columns so that non-label NULLs do not force a full
TSID re-computation.
- Update expected hashes in tests to match the new logic.
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/change-tsid-gen:
test: add comprehensive TSID generation tests for label ordering and null handling
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/change-tsid-gen:
bench: add criterion benchmark for TSID generator
- Compare original mur3 vs current fxhash fast/slow paths
- Test 2, 5, 10 label sets plus null-value slow path
- Add mur3 & criterion dev-deps; register bench target
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/change-tsid-gen:
test: stabilize metric-engine tests by fixing non-deterministic row order
- Add ORDER BY to SELECTs in TTL tests to ensure consistent output
- Update expected __tsid values after hash function change
- Swap expected OTLP metric rows to match new ordering
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
* feat/change-tsid-gen:
refactor: simplify Default impls and remove redundant code
- Replace manual Default for TsidGenerator with derive
- Remove unnecessary into_iter() call
- Simplify Option::unwrap_or_else to unwrap_or
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
---------
Signed-off-by: Lei, HUANG <mrsatangel@gmail.com>
