Skip to main content
POST
Pause run

Authorizations

Authorization
string
header
required

Use a JWT bearer token or a Veryfront API key in the Authorization header.

Path Parameters

run_id
string
required
Minimum string length: 1

Response

Pause requested; the run as the request left it.

run_id
string
required
Minimum string length: 1
id
string
required

The run id. Same value as run_id, which it replaces.

Minimum string length: 1
kind
enum<string>
required
Available options:
agent,
workflow,
task,
eval
status
enum<string>
required
Available options:
pending,
running,
waiting,
completed,
failed,
cancelled
owner
object
required
project_id
string<uuid> | null
required

The project the run belongs to. Null only for platform runs outside a project. With conversation_id, replaces owner.

conversation_id
string<uuid> | null
required
message_id
string<uuid> | null
required
usage
object | null
required

AI usage summed over the billed provider calls; token totals may undercount calls whose provider usage was not captured.

tool_error_count
integer
required
Required range: x >= 0
children_failed
boolean
required
parent_run_id
string | null
required
Minimum string length: 1
root_run_id
string
required
Minimum string length: 1
waiting_reason
enum<string> | null
required
Available options:
tool,
approval,
event,
input,
child_run,
manual_pause,
null
waiting_on
object[] | null
required

Active dependencies while waiting_reason is child_run: every listed run must finish before this run continues. Null otherwise, including runs that waited before this field existed.

waiting
object | null
required

Why the run is waiting. Null unless status is waiting. Replaces waiting_reason and waiting_on.

cancellation
object | null
required

Set once the run was cancelled through POST /runs/{run_id}/cancel; null otherwise.

pause
object | null
required

Set while a pause requested through POST /runs/{run_id}/pause governs the run: requested on a pending or running run, or confirmed as its manual_pause wait. Null otherwise.

title
string | null
required

The custom title set at creation with request.name; null when none was set.

target
string | null
required
workflow_id
string | null
required
schedule_id
string<uuid> | null
required
batch_id
string<uuid> | null
required
runtime_target_kind
enum<string> | null
required
Available options:
main_branch,
environment,
preview_branch,
null
runtime_target_environment_id
string<uuid> | null
required
runtime_target_branch_id
string<uuid> | null
required
error
object | null
required

Why the run failed. OUTPUT_TOO_LARGE carries detail { size_bytes, limit_bytes }, and its message names both sizes.

logs
string | null
required
artifacts
any[]
required
duration_ms
integer | null
required
exit_code
integer | null
required
start_mode
string | null
required
timeout_seconds
integer | null
required

Task runs only; null for other kinds. The execution deadline for the whole task run, in seconds, measured from its first started_at and spanning every attempt. At the deadline the runtime aborts ctx.signal and the run fails with error.code RUN_TIMEOUT. Task code that ignores ctx.signal can keep running inside the runtime process. Default 300; task:eval defaults to 7200.

backoff_limit
integer | null
required

Task runs only; null for other kinds. How many times a failed task run is retried, with exponential backoff, before it fails: at most backoff_limit + 1 attempts. Only two failures are retried: the runtime never started the task (connection refused, or HTTP 503 with the runtime RUNTIME_SHUTTING_DOWN admission refusal), or the task threw the framework RetryableError. Every other failure, and any retry whose backoff would pass the timeout_seconds deadline, fails the run at once. Default 3; task:eval defaults to 0.

input_schema_sha256
string | null
required

Lowercase sha256 hex of the canonical JSON Schema of the declared input schema this run ran against. Null when none is declared or the run predates schema identities.

Pattern: ^[0-9a-f]{64}$
output_schema_sha256
string | null
required

Lowercase sha256 hex of the canonical JSON Schema of the declared output schema this run ran against. Null when none is declared or the run predates schema identities.

Pattern: ^[0-9a-f]{64}$
input_schema_hash
string | null
required

The declared input schema this run ran against, as sha256:. Replaces input_schema_sha256.

Pattern: ^sha256:[0-9a-f]{64}$
output_schema_hash
string | null
required

The declared output schema this run ran against, as sha256:. Replaces output_schema_sha256.

Pattern: ^sha256:[0-9a-f]{64}$
trigger
object | null
required

How and by whom the run was started. Null for runs recorded without a trigger principal. Replaces trigger_kind, trigger_id and trigger_principal_*.

trigger_kind
enum<string> | null
required
Available options:
manual,
schedule,
webhook,
api,
null
trigger_id
string | null
required
created_by
string | null
required
Required string length: 1 - 255
updated_at
string
required
created_at
string
required
started_at
string | null
required
completed_at
string | null
required
metadata
any
control
object
input
any
config
any
output
any

The final output of a completed run. Its JSON serialization is at most 1,048,576 bytes (1 MiB) of UTF-8. A larger output is never stored or truncated: the run fails with error.code OUTPUT_TOO_LARGE and output null.

trigger_principal_type
enum<string> | null
Available options:
system,
user,
api_key,
service_account,
null
trigger_principal_id
string | null