Operations

Easy BPM includes operational tools for investigating process execution, recovering failed work, and cleaning old runtime data.

Instance timeline

The Admin Console Instance Explorer shows a chronological timeline for each process instance. The timeline is different from nodeHistory:

View	Purpose
Node history	Shows the path of process nodes visited by the instance.
Process timeline	Shows timestamped runtime events such as task creation, task completion, worker requests, incidents, retries, and manual moves.

Timeline events are useful for support investigations because they show both process movement and operational actions.

Common event types include:

Event type	Meaning
PROCESS_STARTED	Instance was created and started.
NODE_ENTERED	Execution entered a node.
TASK_CREATED	A user task was created.
TASK_CLAIMED	A user claimed a task.
TASK_COMPLETED	A user completed a task.
WORKER_REQUESTED	An API or external service task was sent to the worker.
WORKER_COMPLETED	Worker completed the external task.
WORKER_FAILED	Worker task failed or timed out.
GATEWAY_EVALUATED	A gateway routed execution to one or more next nodes.
INCIDENT_CREATED	An operational incident was created.
INCIDENT_RETRY_REQUESTED	An operator requested a retry from Incident Manager.
INCIDENT_RESOLVED	An incident was marked resolved.
MANUAL_MOVE	An operator moved the process token.
PROCESS_COMPLETED	Instance completed.
PROCESS_FAILED	Instance failed.
PROCESS_CANCELLED	Instance was cancelled.

Incident management

Incidents are operational records created when Easy BPM cannot continue work normally or when a task-specific failure needs operator attention.

Incident sources are intentionally broad:

Source	Typical cause
PROCESS_ENGINE	Unhandled runtime error, gateway error, or process failure.
WORKER	API/service task timeout, DLQ, or external worker failure.
CODE_TASK	JAR-based code task execution failure.
AI_TASK	AI provider/configuration/retry failure.
MESSAGE	Message-related runtime failure.

Operators can:

• filter incidents by status, source, and instance
• acknowledge an incident
• retry recoverable worker/API incidents
• resolve incidents with a resolution action
• reopen incidents
• open the related process instance
• inspect the incident timeline

Resolution actions include:

Action	Use when
RESOLVED_MANUALLY	Operator handled the issue outside the system.
VARIABLE_FIXED	Process variables were corrected before recovery.
RETRIED_SUCCESSFULLY	A retry fixed the incident.
IGNORED_KNOWN_ISSUE	Issue is accepted or tracked elsewhere.
INSTANCE_CANCELLED	Process instance was cancelled as the resolution.

Purge and archiving

The Admin Console Maintenance page provides destructive cleanup operations for environments with data-retention policies.

Available operations:

Operation	What it deletes
Configured retention policy	Completed process instances and completed tasks that exceed the saved retention windows.
Purge completed instances	Completed process instances older than a selected date, optionally filtered by process definition.
Purge completed tasks	Completed tasks older than a selected date without deleting the parent process instance.
Delete process definition	A process definition and all runtime data related to that definition.

All maintenance operations support a preview before execution.

The configured retention policy also supports:

• enable or disable scheduled cleanup
• separate retention windows for completed process instances and completed tasks
• a batch size limit for each cleanup run
• a Spring cron schedule for automated execution

The cleanup removes related runtime records, including:

• process instances
• tasks
• process variables
• task variables
• documents
• message subscriptions
• worker requests
• code-task execution audits
• incidents and incident events
• process timeline events
• call-activity mappings

Use these operations carefully. They are hard deletes, not soft archive records.

Recommended support workflow

1. Search the process instance in Admin.
2. Review the process timeline and current variables.
3. If an incident exists, open it from Incident Manager.
4. Fix variables or external system state if needed.
5. Retry worker/API incidents where applicable.
6. Resolve the incident with a resolution action.
7. Use purge tools only after the customer retention window has passed.