Test schedules

A schedule combines a project worker, source, artifact rule, check type, cadence, success criteria, and notification behavior. It is the saved configuration used for both recurring and manual checks.

Create a schedule

1. 1

   Choose the worker

   Enter a clear name and select one online project worker. Set the scratch volume size in GiB; it must hold the download, the unpacked dump and the restored database. For a database restore, you can also set CPU and memory requests and limits for the temporary database container. Empty fields leave the restore Job at its default BestEffort behavior.

2. 2

   Select the source and artifact

   Choose a source, then use Automatically latest created file or Filename with date variables. The path is the folder that holds the backups, relative to the source and written with /. Optionally reject artifacts outside a maximum age or minimum/maximum size.

3. 3

   Choose the check type

   Select File or a supported database and exact version from the live catalog. For databases, configure the container, compression, dump format, optional encryption, and target database where available.

4. 4

   Set the cadence

   Choose hourly, daily, or weekly and set the UTC time controls. Hourly schedules are available only when the current plan permits them.

5. 5

   Define success

   Database schedules require validation gates. File schedules always require the selected file to exist and can apply age and size criteria.

6. 6

   Configure notifications

   Choose Never, Always, or On error only, then select active notification channels.

7. 7

   Save and optionally run

   Review the configuration. A new schedule starts active. You can start a manual check immediately without changing its next scheduled run.

Select the intended artifact

Automatically latest created file selects the file with the latest modification time reported by the source, directly in the configured folder; subfolders are not searched. The console label does not mean filesystem creation time. Filename with date variables resolves the folder and the filename in UTC. For example, the folder backups/{{$yyyy}}/{{$mm}} with the filename:

backup_{{$yyyy}}-{{$mm}}-{{$dd}}.sql.gz

Date variables work in both the folder and the filename. Wildcards (*, ?) work only in the filename; a path with a wildcard is rejected when you save the schedule. The console lists all supported variables and arithmetic offsets. You can also tell RestoreTrust to choose the latest modification time when more than one file matches a pattern.

Understand schedule timing

Schedule times and date-variable resolution use UTC, even when the console displays timestamps in your local time. The console's App settings → Time zone choice changes timestamp display and date-and-time inputs for filters and reports, but it does not change a schedule's UTC cadence. A daily UTC time can therefore appear one hour later or earlier locally when daylight saving changes. See Time zones and dates for an example.

For recurring schedules, RestoreTrust chooses one minute within the ten-minute window selected in the console when you save the schedule; do not depend on execution at every minute in the window. A check may also wait in Queued while worker capacity is unavailable. The displayed Next run time is converted to the chosen console zone.

Set database CPU and memory

Database schedules can optionally set independent CPU and memory requests and limits for the temporary database container. A request helps Kubernetes place the Job; a limit caps that container. The other restore containers have no CPU or memory settings. There is no separate BestEffort toggle: leave all four fields blank to create BestEffort restore Pods. File checks do not offer database resources.

Use CPU and memory for workers and Jobs for the field ranges, examples, the agent's separate resource profile, and what happens when a request cannot be scheduled or a memory limit is exceeded.

Pause a schedule to stop future automatic runs, and reactivate it when ready. Existing checks and their results remain separate records. Plan quota, worker availability, and the schedule's active state all affect whether a new recurring check can run.

Choose a check type

• Database restores download, decode, import, health-check, and probe a PostgreSQL, MySQL, MariaDB, or MongoDB backup.
• File checks verify selection, age, size, download, and optional age decryption without importing the content.