Releases: feldera/feldera
Releases · feldera/feldera
v0.8.0
Added
- API: Generalized upsert operation that allows partial modifications to table records (#1296)
- SQL: Functions on binary string (octet_length, position, overlay, substring) (#1264)
- pipeline-manager: add PUT endpoints for Programs, Pipelines, and Connectors (#1248)
- Documentation: Adding a markdown page for videos (#1126)
- SQL: Add hyperbolic functions (#1280)
Fixed
- pipeline-manager: fix a resource usage problem with http streaming under high load (#1257)
- SQL: Changed the semantics of integer arithmetic to match SQL standard (#1247)
- WebConsole: Connector name change doesn't reflect in the pipeline builder (#1238)
- WebConsole: UI sends HTTP request in an infinite loop (#1085)
- pipeline-manager: allow docs website to CORS allowed origins (#1285)
Changed
- Python demo and test scripts are standalone as they no longer depend on dbsp python library (#1230)
- pipeline-manager: update service endpoints to use names instead of IDs in URLs, and add PUT endpoint (#1263)
- SQL: Changed the semantics of integer arithmetic to match SQL standard (#1247)
- pipeline-manager: use names instead of IDs in API endpoints (#1214)
- WebConsole: use names instead of IDs as an API entity identifier (#1214)
- WebConsole: Bearer token expiration now triggers a background token refresh or redirects to a login page (#1100)
- Minimal rust version required to build feldera increased to 1.75 (was 1.73).
v0.7.0
Added
- WebConsole: Add ability to edit connector configuration as JSON
- SQL: Preliminary support for computations with bounded memory on unbounded
streams (#1197)
Fixed
v0.6.0
Fixed
- WebConsole: Vendor logos now change color when in dark mode
Added
- SQL: support for trigonometric functions
sin
andcos
(#1118) - SQL: support for mathematical constant
PI
(#1123) - WebConsole: 'Inspect connector' button in the connector list in Pipeline Builder that opens a non-editable popup (#1128)
- SQL: Support for user-defined functions, declared in SQL and implemented in Rust (#1129)
- SQL: support for other trigonometric functions supported by Calcite (#1127)
- WebConsole: Add Settings page, add a view to manage API keys (#1136)
v0.5.0
v0.4.0
Fixed
- [SQL] Fix bugs in parsing of KEY and FOREIGN KEY constraints
- Use better defaults for running the pipeline-manager (#994) (#1011)
- WebConsole: Fix unable to delete orphaned output connectors,
SQL views in Pipeline Builder do not get removed when removed from the program (#854) - WebConsole: Replace youtube link on Webconsole UI home page (#935)
- WebConsole: Kafka authentication protocol security.protocol field (#963)
- WebConsole: Differentiate compiling and pending in web-console editor (#695)
- WebConsole: Report number of parsing errors per connector (#776)
- WebConsole: Table browser doesn't always correctly apply updates - no longer reproduces (#635)
- WebConsole: MUI DataGridPro fails to load data on Next.js > 13.4.8-canary.9 - no longer reproduces (#494)
- WebConsole: Spacing for SQL icon is off in pipeline view (#932)
- WebConsole: Pipeline failure state in Pipeline Management isn't cleared immediately - no longer reproduces (#1012)
- REST API: Fold ResourceConfig into RuntimeConfig to allow users to configure resources (#1035)
Added
v0.3.2
v0.3.1
v0.3.0
Removed
- Removed support for
FLOAT
SQL data type, since it is ambiguous.REAL
is recommended instead. (#980)
Added
- Support various forms of TopK computations in SQL (#968)
- Support for ORDER BY with LIMIT (#954)
- Allow configuring resource requirements per pipeline (940)
- Restructure and expansion of cloud documentation (#957)
- Secrets can be referenced using a string pattern in the Kafka connector input and output configuration (#949, #970)
- Number of records in the pipeline table in the web console is no longer rounded (#967)
- Ability to pause/resume row updates in Data Inspection tables (#603)
Fixed
- Reduce Docker logging noise from Kafka connect and Redpanda.
- Regression in pipeline shutdown logic (#961)
v0.2.0
Fixed
- Avoid shutting down pipelines when they encounter errors during lifecycle state changes. (#869)
- The compiler flag
-js
generates primary key information (#772) - Fixes a regression caused by using the --auth-provider argument for the pipeline-manager in docker-compose, which was backward incompatible (#900)
Added
- Preliminary support for BINARY and VARBINARY SQL data types (#917)
- Experimental Snowflake sink (#774)
- WebConsole: Snowflake output connector dialog and node (#859)
- Source and sink connector documentation (#882)
- Enforce distinct outputs. This is equivalent to applying
SELECT DISTINCT
to each output view. (#871) - Ignore outermost
ORDER BY
clauses, which don't make sense for incremental queries. (#883)
v0.1.7
Added
- Added a
--lenient
SQL compiler flag to allow views with multiple
columns with the same name. - Added a demo based on the "Feldera: The Basics" tutorial. People who don't
want to manually complete all steps in the tutorial can instead play with the
pre-built pipeline (#822) - Support input tables with primary keys (#826)
- Add health check endpoints to pipeline-manager components (#855)
- Web Console: Add confirmation dialog for delete actions (#766)
- Added documentation for deploying Feldera Cloud on AWS EKS (#850)
- DB migration until now was performed during DB connection setup. Now, users
running the standalone services must invoke the new migrations binary to
explicitly perform database upgrades. The pipeline-manager binary retains the
old behavior for convenience (#856)
Fixed
- Busy-wait loop in Kafka producer (#842)