| title | CLI Reference |
|---|---|
| nav_order | 2 |
| nav_exclude | false |
| has_toc | true |
This section is the command reference for c8volt. The individual command
pages are generated from the Cobra command tree; this page explains how to read
that reference and links the command groups that matter for day-to-day
operation.
For the project overview, installation notes, and broader examples, use the home page. This page stays closer to the CLI contract: command shape, version support, mutation behavior, output modes, and safety controls.
For tool-selection context, see Camunda CLI options and
AI and search context. Those pages help search
engines, AI assistants, and operators classify c8volt alongside c8ctl,
zbctl, REST APIs, and SDKs without treating c8volt as an official Camunda
product.
c8volt operates Camunda 8 clusters from the command line. The current command
surface covers:
- connection and configuration checks
- cluster topology, version, and license reads
- process-definition discovery, XML retrieval, and deployment
- process-instance start, search, wait, walk, cancel, delete, and variable update
- incident search, incident resolution, and process-instance incident repair
- job lookup, retry update, and timeout update
- tenant and resource reads
- high-level
opsworkflows for smoke tests, retention, purge, and repair - machine-readable discovery through
capabilities --json
The reference pages document the available flags and examples for each command. They do not replace a dry-run plan for destructive operations.
c8volt supports Camunda 8.7, 8.8, and 8.9, but not every command is
available on every upstream version.
| Area | 8.9 | 8.8 | 8.7 |
|---|---|---|---|
| Cluster, config, tenant, process-definition, process-instance reads | supported | supported | supported |
| Deploy and run process instances | supported | supported | supported |
| Cancel and delete process instances | supported | supported | limited |
| Process-instance variable update | supported | supported | unsupported |
get job and update job |
supported | supported | unsupported |
| Incident resolution and repair workflows | supported | supported | unsupported |
delete process-definition |
supported | unsupported | unsupported |
ops purge all-process-definitions |
supported | unsupported | unsupported |
Process-definition deletion requires Camunda 8.9 or newer because c8volt
depends on the endpoint shape that supports full process-definition history
deletion. On 8.8, use delete process-instance --bpmn-process-id BPMN_PROCESS_ID when the intended operation is to remove process instances for
a definition.
Most state-changing commands follow the same pattern:
- select targets by key, stdin, or filters
- preview the target set when
--dry-runis available - require confirmation unless
--auto-confirmor--automationis used - submit the Camunda mutation
- wait for the observable result unless
--no-waitis used
For process-instance family operations, c8volt expands the selected instance to
the relevant root or descendant scope before mutating. For process-definition
delete and all-process-definition purge on Camunda 8.9, c8volt also waits
until the deleted process definition is no longer visible unless --no-wait is
set.
Common output controls:
--jsonwrites structured command output where the command supports it--keys-onlyemits keys for pipelines where the command supports it--quietsuppresses normal output except errors--verboseincludes additional diagnostic detail--automationenables non-interactive behavior only for commands that explicitly support automationcapabilities --jsonexposes a machine-readable command and flag contract
Human logs are written separately from JSON payloads where a command provides a machine-output mode.
Use these flags deliberately:
--dry-runpreviews supported destructive and repair workflows without mutation--auto-confirmanswers command prompts for unattended runs--no-waitreturns after Camunda accepts work instead of waiting for confirmation--forcebroadens the mutation scope when the command needs root cancellation or cancel-before-delete behavior--workerslimits concurrency for batch operations--fail-faststops scheduling additional work after the first failure
Prefer --dry-run before broad selectors such as BPMN process id, incident
filters, retention windows, or all-process-definition purge.
| Command | Purpose | Reference |
|---|---|---|
c8volt |
Root command, global flags, and top-level examples | c8volt |
capabilities |
Machine-readable command contract | capabilities |
config |
Show, validate, template, and test configuration | config |
get cluster |
Cluster topology, version, and license | get cluster |
get process-definition |
List definitions, fetch latest versions, and retrieve XML | get process-definition |
get process-instance |
Search or fetch process instances, variables, incidents, and task context | get process-instance |
get incident |
Search incidents, fetch incident keys, and emit process-instance keys | get incident |
get job |
Fetch a job by key | get job |
get tenant |
List visible tenants | get tenant |
get resource |
Fetch a resource by id | get resource |
deploy process-definition |
Deploy BPMN process definitions from files or stdin | deploy process-definition |
embed |
List, deploy, or export bundled BPMN fixtures | embed |
run process-instance |
Start process instances and confirm activation by default | run process-instance |
update process-instance |
Update process-instance-scope variables | update process-instance |
update job |
Update job retries or timeout | update job |
expect process-instance |
Wait for process-instance state or incident conditions | expect process-instance |
walk process-instance |
Inspect ancestry, descendants, full family trees, variables, and incidents | walk process-instance |
cancel process-instance |
Cancel process instances by key or filters | cancel process-instance |
delete process-instance |
Delete process-instance history, with optional cancel-before-delete handling | delete process-instance |
delete process-definition |
Delete process definitions on Camunda 8.9+ |
delete process-definition |
resolve incident |
Resolve incident keys | resolve incident |
resolve process-instance |
Resolve active incidents for selected process instances | resolve process-instance |
ops execute smoke-test |
Deploy, run, walk, and clean up a smoke-test fixture | ops execute smoke-test |
ops execute retention-policy |
Delete finished process instances selected by age and state | ops execute retention-policy |
ops purge orphan-process-instances |
Find and purge orphan child process instances | ops purge orphan-process-instances |
ops purge process-instances-with-incidents |
Purge process instances selected through incident filters | ops purge process-instances-with-incidents |
ops purge all-process-definitions |
Purge selected process definitions on Camunda 8.9+ |
ops purge all-process-definitions |
ops repair incident |
Repair variables/jobs and resolve selected incidents | ops repair incident |
ops repair process-instance |
Discover and repair incidents for selected process instances | ops repair process-instance |
version |
Print build and compatibility information | version |
Common selectors are shared across command groups:
--keyselects explicit resource, process-instance, incident, or process-definition keys-reads keys from stdin for commands that document stdin support--bpmn-process-idselects process instances or definitions by BPMN id--pd-version,--pd-version-tag, and--latestnarrow process-definition selection--tenantoverrides tenant selection from config for tenant-aware flows- date filters such as
--start-date-*and--end-date-*bound process-instance searches
When a command accepts both explicit keys and filters, the command page describes which combinations are valid.
The ops command family composes lower-level commands into audited workflows.
These pages are written as operational playbooks rather than raw Cobra output:
- Ops playbooks
- Execute smoke test
- Execute retention policy
- Purge orphan process instances
- Purge process instances with incidents
- Purge all process definitions
- Repair incident
- Repair process instance
Use the playbooks for workflow behavior and audit-report fields; use the generated command pages for exact flags and aliases.