JMA

Job Management

20260115.200 (API v9.1.0)

• Allow filtering ProcessingSteps by ProCon version property, by exact string match

20260113.195 (API v9.0.0)

• Allow filtering and sorting by versions
• Allow version ranges to be used (similar to npm)
• Fix issue with admin query: public ProcessingSteps was not visible when admin was using normal query
• More actions now require code contributor permission
• Reworked registering of PS, now only an upload is used and version, function name are taken from manifest
• ProcessingSteps now contain the ProCon version used when a deployment exists
• ProcessingStep property edit is more limited now
• Platform user and owner renamed

20251209.186 (API v8.8.0)

• Fix copy PS to user and org did not return new PS ID
• Add Action classes “Destructive” and “LongRunningBackground” for more information on the effect
• Add fix OTL metric requested non existing PS Queues if the PS was deprecated
• Do reverse mapping of grants and roles in info endpoint
• Add remote endpoint in info HTO which is used to connect with remote workers Info->Remote
• Now all JMA actions have modern Draft 2020-12 JsonSchemas
  • more properties are now required in the schema to allow the caller to check early

20251202.175 (API v8.8.0)

• Processing steps can now be deployed directly via JMA API.
• Registering or deploying a ProcessingStep now requires CodeContributor permissions
• A CodeHash must be assigned beforehand to any processing step that is to be deployed. Existing CodeHash cannot be changed but it can be removed by an admin and assigned again by a user with sufficient authority.
• Afterwards, the processing step can be deployed by optionally configuring these parameters or using default values: ResourcePreset, EntryPoint and Scaling.
• A deployed step can be re-deployed with a different configuration infinite number of times.
• A deployed step may be suspended by the user, in which case the configuration is saved but the deployment itself is removed. A suspended deployment can be resumed later with the saved configuration.
• A processing step that is not platform deployed can optionally be marked as “externally deployed” to provide information that it is externally deployed.
• Deployments can be removed at any time by the user.

20251117.161 (API v8.6.0)

• Use the new background worker feature for “restore all job read models” and “restore some job read models” operation.

20251104.151 (API v8.6.0)

• Add API events, provides an SSE feed with information on changes
• Add cache headers to EntryPoint, and Root nodes and parameter schemas so clients know they can be cached
• Allow caching of CORS response
• New RMQ queue layout, including the organization context
  • no function naming collision between contexts
• Present registry and SSE endpoint in Info
• Fix action titles do now contain a title usable in UI

2025.10.27.142 (API v8.5.0)

• Processing steps can now be deprecated.
  • Jobs can not start with deprecated function
  • No RMQ queues maintained for deprecated functions

2025.09.08.131 (API v8.3.0)

• Separate admin and normal query for used tags for Jobs, steps and WorkDatas.

2025.08.28.128 (API v8.2.0)

• Add action to delete Processing step. It only succeeds if the step is not used by any job.

2025.08.18.127 (API v8.1.0)

• Add actions to copy WorkData and Processing steps from organization to user and conversely from user to organization.

2025.07.28.120 (API v8.0.0)

• Any variable names, function parameter names, header keys, etc. related to PermissionGroups renamed to PermissionGrants.
• The OwnerId property of Job, WorkData and ProcessingStep entities is now split into two properties: OwnerId and CreatedBy. CreatedBy is the Id of User who created this resource. OwnerId is the Id of the organization if the user was in an organization context during this entity’s creation, else the user is considered to be in a private context and it’s value is same as CreatedBy.

2025.07.25.112 (API v7.6.0)

• Add function and version naming conventions
  • functions: a-Z, 0-1, '', optional leading '' is allowed
  • versions: a-Z, 0-9, ’.’, ’_’ ’-’
• Ensure versions do not interfere with routing
• Cleanup communication

2025.07.25.107 (API v7.6.0)

• Fix JMA version was not show in API

20250521.67 (API v7.6.0)

• Improved performance of rapid job setup

20250514.65 (API v7.6.0)

• Add action to create and configure job with a single HTTP request
• This new action allows specification of parameters to create a new job, select processing step, configure processing parameters, set tags, toggle whether to output data deletion, set input data slots, and finally start the job.

20250506.55 (API v7.5.0)

• Add action to hide processing steps
• Add ability to query Processing steps by IsConfigured parameter
• Job can now show its started time if it has been started
• WorkData filename is also passed in JobOffer message

20250409.1 (API v7.4.0)

• Fix job could get stuck in pending state if it could not be scheduled internally. Now job will transition to error.

20250403.4 (API v7.4.0)

• Streamed file upload: if a file is big it will now be processed in chunks
• Added owner field to WorkData, Job, ProcessingStep
• Fixed issue with admin queries

20250214.1 (API v7.3.1)

• Improve security model by passing signed credentials to ProCon for accessing the API
• Homogenize the URL building so ProCon uses same URLs

20241212.1 (API v7.3.1)

• Fix missing Name property for DataSlots in Job

20241127.4 (API v7.3.0)

• Add additional navigation links to the relevant WorkData: “Job that produced this work data” and “Processing step that produced this work data”
• Expand work data query with these additional filter filters:
  • ProducerProcessingStep (Filters WorkData that are produced by the processing step represented by this URL)
  • IsUsed (Filters WorkData by the condition if any job is using them)
• Expand job query with an additional filter parameter:
  • IsUsingThisWorkData (Filters jobs that are using the work data represented by this URL)
• Fixed a bug where some collection data slots were not marked as collection. IsCollection property is now correctly derived from min and max files.

20241118.3 (API v7.2.0)

• Extend ProcessingStep registration and editing to allow versions to be set and edited
• ProcessingStep versions are now used every where
• Fix rename of function did not update internal communication so messages

20241112.1 (API v7.1.0)

• Internal preparation for workflows

20241104.1 (API v7.1.0)

• Add configured parameters to jobs so it is possible to see what is / was configured
• Allow jobs to run with default parameters without explicitly resetting them

20241018.2 (API v6.2.4)

• added new admin features manage broken jobs
• added new filter for jobs createdBefore and createdAfter

20241016.4 (API v6.2.3)

• improved resilience
• transient storage errors are communicated in 503 with retry-after suggestion in more cases
• stability fixes

20241009.4 (API v6.2.3)

• Added processing step function versioning capability.
• Validation are set on function name and version. User is prompted while registering if the processing step with same function name and version exist in the DB.
• The Default value without user input is set to “0”, else user is allowed to input 50 characters for version and for processing step function name limit is 170 chars.
• Delete Operation on WorkData is now only shown, if it can be executed (internal check if it is used)
• Performance fixes and better error messages

20240919.5 (API v6.2.2)

• Fix issue that enables to work with versioned and non-versioned bucket for MinIO, Azure Blob Storage and S3 Binary Storage

20240619.3 (API v6.2.1)

• Fix output WorkData processing after job completion (size is updated, empty ones are deleted)
• Fix internal worker communication to API by adding configuration options for internal API host URL

20240618.1 (API v6.2.0)

• Added a new boolean condition called ‘IncludeRemainingTags’ to JobReadModel, WorkData, and ProcessingSteps queries. This will allow us to include the remaining tags for filtered objects.
• Added a feature that enables us to retrieve the list of RemainingTags for JobReadModel, WorkData, and ProcessingSteps.

20240617.1 (API v6.1.0)

• Added Job deletion feature. A job can be deleted if certain conditions are met.
• The deletion of job and any related data is permanent. Only the events log is retained for future reference.
• Once the job is deleted, existence of any entity referred in the event log cannot be guaranteed.

20240523.1 (API v6.0.0)

• WorkData for a job’s input/output DataSlots can now be collections. Configure DataSlots with WorkData now accepts multiple WorkData
• WorkData filenames now can only consist of ASCII characters and are limited to 1024 chars
• Stability fixes for DB access (retries)
• Stability fixes in communication with Workers (retry, reconnect)
• Rename of inconsistent Property in Job CreatedOn-> CreatedAt

20240409.1 (API v5.1.1)

• Fix issue where jobs could stick in pending state
• Fix issue with unordered events causing incorrect job states

20240402.1 (API v5.1.0)

• Info endpoint now contains used bytes in storage
• Added tag query for WorkData. Can be restricted by required tags which will return all tags of WorkData which exist next to the required tags

20240319.1

• Moved User id to string (from UUID)

20240312.1

• Fix enum name mapping incorrect for sort property
• Fix WorkData accepts now empty string as comment

20240311.1

• Fix pagination issue where property name in enum was not matching real property
• Add PreviousEntitiesCount to query results so it is known what entities are shown

20240305.1

Add capabilities to process meta data send from ProCon

20240304.2

• Fix JSON schema validation error.
• Better JSON validation errors in API responses

20240129.1

• Fixed signing presigned URL, uploads now work properly

20240209.1

• WorkData “mark as deleted” is renamed to “hide”
• WorkData that is not used as input for a Job can now be marked as deletable and be deleted
• Jobs can mark all generated data as “deletable”