CHANGELOG.md

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

0.40.0 - 2024-03-27

Changed

• • BREAKING: remove type from datamanager table (#394)
• • [chore] towncrier is now used for changelog management (#395)

0.39.0 - 2024-03-07

Changed

• Compute task status DOING is renamed EXECUTING (#371)

0.38.0 - 2024-02-26

Removed

• BREAKING: remove all code related to the distributed mode, and mentions in schemas and documentation (#341)
• BREAKING: distributed Skaffold profile and mentions in doc (#319)
• BREAKING: chaincode-init and chaincode Dockerfiles (#319)
• Flag & environment variables to choose between standalone and distributed mode (#347)

Added

• Add Image addressable to Function object. By default, Image is set to an empty string adressable. This addressable is updated when the checksum and storageAddress are available (#288)
• Enum FailedAssetKind (#277)
• BREAKING: Field asset_type of type FailedAssetKind in FailureReport (#277)
• BREAKING: Add FunctionStatus (#263)
• Add Function status event machine (#263)
• BREAKING: Add statuses WAITING_FOR_BUILDER_SLOT and BUILDING on tasks to reflect associated function status (#366)
• Add task actions BUILD_STARTED and BUILD_FINISHED to propagate status change from function to compute task (#366)

Changed

• Rename Function addressable to Archive (#288)
• Renamed compute_task_keyby asset_key in FailureReport (#277)
• FailureReport now can be reference a ComputeTask or a Function through asset_key + asset_type (#277)
• Logic to determine new compute task status takes in account the status of the function. A new task can now be created with the status FAILEDor CANCELLED (if the function reached the corresponding status) (#365)
• BREAKING: Transition to status TODO for a given compute task is done after the function is built(#365)
• BREAKING: Rename TODO to WAITING_FOR_EXECUTOR_SLOT and WAITING to WAITING_FOR_PARENT_TASKS(#366)

Fixed

• incorrect link in documentation (#307)
• Skip compute ask permissions checks when the action is propagated from the function status change (#367)

0.37.0 - 2023-10-18

Added

• Source code formatting (#313)

Changed

• Bump go version to 1.21 (#316)

0.36.1 - 2023-10-06

Added

• three_orgs Skaffold profile for standalone orchestrator (#280)

0.36.0 - 2023-09-07

Changed

• Replace deprecated ioutil package with os (#269)

0.35.2 - 2023-07-25

Changed

• Minor dependency updates. See commit history for more details.

0.35.1 - 2023-06-27

Changed

• Minor dependency updates. See commit history for more details.

0.35.0 - 2023-06-12

• No changes on the app

0.34.0 - 2023-05-11

Changed

• A Performance in now unique regarding a compute task key, a metric key and a compute task output identifier (#197)

Removed

• Metric from Performance (#213)

0.33.0 - 2023-03-31

Changed

• BREAKING: rename Algo to Function (#139)

0.32.0 - 2023-01-31

Added

• Context in fsm calls (#127)
• Contributing, contributors & code of conduct files (#123)

Removed

• Test Only field for data samples (#116)

0.31.1 - 2023-01-09

Removed

• field parent_task_keys in ComputeTask (#112)
• view expanded_compute_tasks (#112)

[0.31.0] - 2022-12-19

Changed

• bump app version to 0.31.0

Fixed

• end-2-end postgres dump upload (server version)

[0.30.0] - 2022-11-22

Removed

• (BREAKING) task category
• (BREAKING) task specific data

Changed

• TASK_UNKNOWN is a valid category
• Allow registration of performances on any task category
• Update the TLS certificates (#91)

[0.29.0] - 2022-10-03

Added

• allow setting gRPC keepalive enforcement policy

Removed

• (BREAKING) delete_intermediary_models property in ComputePlan and NewComputePlan
• (BREAKING) ModelCategory and associated Model & NewModel fields
• (BREAKING) AlgoCategory and associated Algo and NewAlgo fields

Changed

• failure reports: Build errors now have a logs address
• (BREAKING): Replaced algo by algo_key in ComputeTask

Fixed

• 000042_reference_task_outputs use compute_task_key instead of asset_key for assetKey

[0.28.0] - 2022-09-26

WARNING: Some migrations in this version are destructive once applied you will not be able to restore algo categories.

Changed

• (BREAKING) Algo.category: do not rely on categories anymore, all algo categories will be returned as UNKNOWN
• NewAlgo.category: No category is expected
• (BREAKING): Replaced algo by algo_key in ComputeTask

Fixed

• Task outputs events come after their referenced asset creation event

Removed

• Worker field on NewAggregateTrainTaskData, use NewComputeTask.Worker field instead
• DisableModel rpc in Model service, use DisableOutput in ComputeTask service instead
• CanDisableModel rpc in Model service.

Added

• failure_date field to ComputePlan protocol buffer schema
• IsComputePlanRunning gRPC method in ComputePlanService

Removed

• (BREAKING) compute plan status

[0.27.0] - 2022-09-19

Added

• Worker field on NewComputeTask mandatory for tasks without input data

Deprecated

• NewAggregateTrainTaskData.worker: use NewComputeTask.Worker field instead

Removed

• NewComputeTask.parent_task_keys which was deprecated since 0.26.0
• Restriction on algo-task category matching
• Restriction on parent task category

[0.26.1] - 2022-09-12

Changed

• Images are built using protoc-gen-go v1.18.1, grpc_health_probe v0.4.12 and migrate v1.28.1

Fixed

• Disable output RPC on distributed mode

[0.26.0] - 2022-09-07

Removed

• (BREAKING) ModelService.GetComputeTaskInputModels, use ComputeTaskAPI.GetInputAssets instead
• Test task rank special case: rank is not inherited from parent task anymore
• (BREAKING) QueryModels from model service: it was unused and model category will soon be deprecated

Deprecated

• NewComputeTask.parent_task_keys is deprecated since parent tasks are determined from task inputs

[0.25.0] - 2022-08-29

Added

• New RPC to disable task outputs

Deprecated

• ModelService.GetComputeTaskInputModels, use ComputeTaskAPI.GetInputAssets instead

[0.24.0] - 2022-08-22

Fixed

• Properly register compute task outputs in distributed mode

Removed

• (BREAKING) Remove RabbitMQ

[0.23.0] - 2022-08-17

Added

• New service methods to update algo, compute_plan and data manager name
• gRPC method to get task input assets

Changed

• Prevent duplicate model registration based on task output definition
• Switched to zerolog logging library

[0.22.0] - 2022-08-09

Added

• Build images with go 1.19
• Add a Transient field to the task inputs
• Return an error in distributed mode if a stored event has invalid event or asset kind
• Associate asset with task output on registration

Removed

• Task counts by status from ComputePlan responses

[0.21.0] - 2022-08-01

Added

• Introduce gRPC SubscribeToEvents method in distributed mode

Changed

• Validate task inputs

Fixed

• In standalone mode, lock the events table when inserting events to prevent missing events in SubscribeToEvents gRPC stream

Removed

• Category filter from QueryAlgos rpc
• Legacy compute task permission fields

[0.20.0] - 2022-07-25

Added

• Introduce gRPC SubscribeToEvents method in standalone mode
• Dispatch updated asset event on ComputePlan cancellation

Removed

• Automatic transition to DONE when registering models or performances.

Changed

• updated grpc healthprobe to 0.4.11 in server image
• updated rabbitmq/amqp091-go lib to 1.4.0

Fixed

• properly ignore mocks when building image locally

[0.19.1] - 2022-07-13

Fixed

• SQL query for organization with null address

[0.19.0] - 2022-07-11

Added

• Organization hostname in the organization object
• CancelationDate in the compute plan object

Fixed

• SQL logging was enabled when METRICS_ENABLED flag was passed instead of documented LOG_SQL_VERBOSE
• Prevent disabling model if task has only predict or test children
• Don't timeout when canceling a compute plan

[0.18.0] - 2022-07-05

Removed

• Metadata set in events

Changed

• (BREAKING) Removed the MetricKeys property of test tasks in favor of the generic Algo field

[0.17.0] - 2022-06-20

Added

• Enable transition to DONE through ApplyTaskAction

[0.16.0] - 2022-06-14

Changed

• (BREAKING) rename node to organization

Fixed

• allow a worker to cancel a task it does not own

[0.15.0] - 2022-06-07

Added

• Introduce Predict task type
• Introduce compute task outputs

Changed

• use go test to run e2e tests

[0.14.0] - 2022-05-31

Added

• Introduce empty compute plan status

Changed

• base docker image from alpine 3.15 to alpine 3.16

Fixed

• event asset migration

Changed

• only update status on task update.

[0.13.2] - 2022-05-24

Fixed

• conn busy error when querying Tasks

[0.13.1] - 2022-05-24

Fixed

• conn busy error when querying Algos

[0.13.0] - 2022-05-23

Fixed

• In standalone mode, truncate TimeService time to microsecond resolution to match PostgreSQL timestamp resolution.

Changed

• Disable CGO.
• More validation of Algo inputs (data managers / data samples)

Added

• Introduce compute task inputs existing tasks won't have any inputs
• Embed historical assets in the event messages.

[0.12.0] - 2022-05-16

Added

• New mandatory name field to compute plan

Changed

• Remove event column

[0.11.0] - 2022-05-09

Added

• Add a new ALGO_PREDICT algo category

Changed

• Validate algo inputs and outputs

[0.10.0] - 2022-05-03

Changed

• Remove model asset column
• Remove performance asset column
• Remove datamanager asset column
• Remove datasample asset column
• Algos now have Inputs and Outputs
• The orchestrator-server doesn't run DB migrations on startup anymore

Removed

• ASSET_METRIC kind

[0.9.2] - 2022-04-15

Changed

• Build with go 1.18

Fixed

• Update failure report asset column migration to prevent null value error when migrating a populated database
• Parent tasks keys format validation

[0.9.1] - 2022-04-13

Fixed

• Order parent tasks keys by task position

[0.9.0] - 2022-04-13

Added

• Added ALGO_METRICS Algo category

Changed

• Remove compute task asset column
• QueryAlgos filter "Category" is now "Categories"
• Remove failure report asset column

Removed

• Metrics gRPC routes. Use Algo gRPC routes and ALGO_METRICS category instead.

[0.8.0] - 2022-04-11

Added

• Allow querying datasamples by keys

Changed

• Remove node asset column
• Remove algo asset column
• Remove compute plan asset column

Fixed

• Do not panic on nil filter

[0.7.0] - 2022-03-29

Added

• Expose gRPC metrics
• Expose database transaction and events metrics
• Expose task metrics

Changed

• Log SQL errors regardless of log level
• Remove asset column of nodes table

Fixed

• Publish events sequentially, preserving the order

[0.6.1] - 2022-03-01

Added

• add support for graceful shutdown on SIGTERM signal

Changed

• removed codegen layer and implicit protojson serialization

Fixed

• Cancel all tasks when cancelling a compute plan
• Check for compute plan existence on task registration
• Disallow registration of tasks on a compute plan you don't own

[0.6.0] - 2022-02-18

Added

• add Start and End timestamp filters for EventQueryFilter
• support composite tasks with two composite parents
• Add migration logs
• add owner field to failure report asset
• Add a new endpoint to register multiple models at the same time

Changed

• return datasamples list in RegisterDataSamplesResponse
• return tasks list in RegisterTasksResponse
• store the error type of a failed compute task in a failure report instead of an event
• improve performance of compute_tasks SQL indexes by using dedicated columns instead of JSONB
• improve performance of compute plan queries by leveraging a specific index for status count
• isolation level of read-only queries in standalone mode is now READ COMMITTED
• improve performance of model SQL indexes by using dedicated columns instead of JSONB

Fixed

• set the correct name of the RegisterFailureReport service method used in distributed mode
• Return the correct models in GetComputeTaskInputModels for composite tasks
• timestamp comparison when performing event sorting and filtering in PostgreSQL
• ComputePlan query now uses correct SQL indexes
• Incorrect sort order when checking parent task compatibility

Deprecated

• RegisterModel gRPC method

[0.5.0] - 2022-01-16

Added

• add a logs_permission field to the Dataset asset
• add a GetDataSample method to the DataSample service

[0.4.0] - 2022-01-05

Added

• add filter for compute plan query
• chaincode now properly propagate request ID in every logs
• log events as JSON
• add FailureReport asset to store compute task failure information

[0.3.0] - 2021-11-30

Added

• sort queried events
• expose basic metrics from server, chaincode and forwarder behind METRICS_ENABLED feature flag
• filter queried events on metadata

[0.2.0] - 2021-11-02

Changed

• (BREAKING) Replace objective by metric
• (BREAKING) Multiple metrics and performances per test task
• fail gRPC healthcheck and stop serving on message broker disconnection

Added

• Get task counts grouped by status when querying compute plans

Fixed

• Events queried from the gRPC API now have their channel properly set
• Leverage asset_key index when querying events

[0.1.0] - 2021-10-04

Fixed

• Stable sorting of tasks

Added

• Expose the orchestrator version and chaincode version

[0.0.2] - 2021-09-16

Added

• Expose worker in task event metadata
• Assets expose a creation date
• Query algo by compute plan
• Handle event backlog
• Retry on fabric timeout
• Add request ID to log context

Changed

• Do not retry on assets out of sync
• Do not compute plan status on model deletion
• Reuse gateway connection in distributed mode
• Replace readinessProbe by startupProbe
• Do not cascade canceled status

Fixed

• Properly retry on postgres' serialization error
• Filtering events by asset in distributed mode
• Input models for composite child of aggregate

[0.0.1] - 2021-06-29

Added

• Automatic generation of graphviz documentation from *.proto file definition
• asset management
• asset event dispatch
• standalone database (postgresql) support
• distributed ledger (hyperledger-fabric) support