Releases: kamu-data/kamu-cli
Releases · kamu-data/kamu-cli
Release v0.204.4
[0.204.4] - 2024-09-30
Changed
- CLI command tweaks:
- Make
--yes / -y
flag global - Add confirmation step to
system compact
command - Add support for patterns to
system compact
to process multiple datasets at once - Fixed argument parsing error in
kamu system compact
command
- Make
- Simplified organization of startup initialization code over different components
Fixed
- Broken catalog issue for server and transactional modes
- Added several E2E tests (happy paths) covering the Flows tab in the UI
- Corrected behavior of
MySqlAccountRepository::get_accounts_by_ids()
, for the case of empty IDs collection
Release v0.204.3
[0.204.3] - 2024-09-26
Fixed
- Dataset creation with unique alias but with existing id for FS dataset storage mode
kamu init
: fixed regression in case of usingexists_ok
flag... finally
Release v0.204.2
[0.204.2] - 2024-09-26
Fixed
kamu init
: fixed regression in case of usingexists_ok
flag
Release v0.204.1
[0.204.1] - 2024-09-25
Fixed
- Fixed build regression, in case
web-ui
feature flag is used
Release v0.203.1
[0.203.1] - 2024-09-24
Added
- Added database migration & scripting to create an application user with restricted permissions
kamu delete
command will respect dependency graph ordering allowing to delete multiple datasets without encountering dangling reference
Release v0.203.0
[0.203.0] - 2024-09-22
Added
- Support
List
andStruct
arrow types injson
andjson-aoa
encodings
Release v0.202.1
[0.202.1] - 2024-09-20
Fixed
- Open Telemetry integration fixes
Release v0.202.0
[0.202.0] - 2024-09-20
Changed
- Major dependency upgrades:
- DataFusion 42
- HTTP stack v.1
- Axum 0.7
- latest AWS SDK
- latest versions of all remaining libs we depend on
- Outbox refactoring towards true parallelism via Tokio spaned tasks instead of futures
Fixed
- Failed flows should still propagate
finishedAt
time - Eliminate span.enter, replaced with instrument everywhere
Release v0.201.0
[0.201.0] - 2024-09-18
Added
- REST API: New
/verify
endpoint allows verification of query commitment as per documentation (#831)
Changed
- Outbox main loop was revised to minimize the number of transactions:
- split outbox into planner and consumption jobs components
- planner analyzes current state and loads bunch of unprocessed messages within a 1 transaction only
- consumption jobs invoke consumers and detect their failures
- Detecting concurrent modifications in flow and task event stores
- Improved and cleaned handling of flow abortions at different stages of processing
- Revised implementation of flow scheduling to avoid in-memory time wheel:
- recording
FlowEventScheduledForActivation
event (previously, placement moment into the time wheel) - replaced binary heap based time wheel operations with event store queries
- Postgres/SQlite event stores additionally track activation time for the waiting flows
- in-memory event store keeps prepared map-based lookup structures for activation time
- recording
Release v0.200.0
[0.200.0] - 2024-09-13
Added
- Added first integration of Prometheus metrics starting with Outbox
- Added
--metrics
CLI flag that will dump metrics into a file after command execution
Changed
- Telemetry improvements:
- Improved data collected around transactional code
- Revised associating span objects with large JSON structures such as messages
- Suppressed several noisy, but not very useful events
- Improved Outbox stability when message consumers fail
- Similarly, Task Executor keeps executing next tasks in case running a task results in an internal error