--- title: "Issues" description: "Why Miget stopped an app, a worker, an add-on, or a cron run - and the history behind it." --- When Miget stops an application, a [worker](/deployments/workers), a database or another add-on, or when a [cron job](/deployments/cron-jobs) run fails, it records why. The record stays after the thing runs again, so it is also a history of what has gone wrong. Open the **Issues** page from the dashboard sidebar, or an application, service, or resource's own page, which shows its own latest issues. The same data is at `GET /api/v1/issues`. ## Why something stopped | Kind | Meaning | |---|---| | `out_of_capacity` | Its resource had no room left for it. | | `crash_loop` | It kept crashing, and Miget stopped it. | | `out_of_memory` | It ran out of memory after every restart, and Miget stopped it. | | `awaiting_checkout` | A [worker](/deployments/workers#free-compute-and-held-workers) of a free-compute application - free compute covers the web process alone - held until the workspace starts [Usage Billing](/pricing/usage-billing#the-monthly-minimum). | | `cronjob_failed` | A [cron job](/deployments/cron-jobs#run-history)'s last run failed. It keeps its schedule; the next successful run resolves the issue. | An issue opens when the condition starts and resolves when it ends - a worker or a database is not started again by a retry on your part: a worker comes back with the next deploy, and an application, service, or add-on starts again once its cause is fixed (more memory, room on its resource, or Usage Billing started). ## Filtering In the dashboard, open an application, service, or resource's own page to see only its issues. Over the API, filter `GET /api/v1/issues` with one of: - `app_id` - an application and everything under it: its own issues, its workers', and its cron jobs' - `service_id` - a standalone service - `resource_id` - a resource, and everything running on it that you can see They are mutually exclusive - sending more than one is a `400`. Leaving all three out does **not** return every issue you can see. It is narrowed to whichever of `apps:view`, `services:view`, and `resources:view` you hold, one kind at a time - a role with only `apps:view` sees application, worker, and cron job issues but no service or resource ones, with no error to say so. ## What an issue carries - `subject` - what stopped: its `type` (`app`, `worker`, `addon`, `service`, or `cronjob`), `uuid`, `name`, and `label` - `open` - true until it is resolved - `opened_at` / `resolved_at` - `details` - specific to the kind: requested and allowed replicas for `out_of_capacity`; restart count (and, for an add-on, the exit code and the last lines of its log) for a crash; a failed run's job name, reason, exit code, and message for `cronjob_failed` ## Next steps Get notified the moment something stops, instead of polling Issues. Where a worker's own issues come from.