Skip to content

List saved job concurrency limits

GET
/admin/jobs/concurrency

Returns the saved concurrency limits for background job types.

You need to be signed in with system.jobs.read or system.admin for the whole install; a permission held only for a team doesn’t count.

The list combines the handler types registered in this process with type names that have a positive, readable saved limit, sorted by name and not paginated. A registered type with no such setting has cap zero. An unregistered name without a positive setting is left out, and if both sets are empty there are no rows.

These are saved settings, not the limits running workers currently enforce. Each application process loads them when it starts. If that read succeeds, positive limits apply within that process’s worker pool, and the pool’s overall worker count still applies. The limits aren’t an install-wide ceiling and don’t constrain claims made through the external-worker API. If the startup read fails, the failure is logged and the pool runs without per-type limits.

applies_on_restart is always true. Changing a setting doesn’t update existing pools. Missing, non-positive or non-integer values are treated as unset, and a positive stored value isn’t clamped to the write endpoint’s maximum of 64 when read. The endpoint reads settings straight from the database, with no configuration cache.

Saved type limits and the restart requirement, not the limits currently active in running pools.

Saved job-type limits, combined with handlers registered in this process.

object
items
required

Registered handler types plus unregistered names with positive, readable saved limits, sorted by name. Registered types without such a setting have cap zero.

Array<object>

A saved per-type limit for in-process workers, not a report of a running pool’s active limit.

object
type
required

Registered job type or an unregistered name with a positive saved limit.

string
cap
required

Saved limit on concurrent runs of this type within each application process’s worker pool, after a successful startup load. Zero means no per-type limit, though the overall pool size still applies. External-worker claims aren’t covered.

integer
applies_on_restart
required

Always true. Each process reads limits at startup; saving a value does not update existing pools, and a failed startup read leaves that pool without per-type limits.

boolean

Authentication required, missing, or invalid

object
error
required

Human-readable error summary

string
Example
the request could not be completed
Example
{
"error": "authentication required: sign in and retry with a valid session or API token"
}

Authenticated but missing required capabilities

object
error
required

Human-readable error summary

string
Example
the request could not be completed