Duplicate schedule detection

Find jobs with identical schedules so you can spread their runs out. Cronstable groups jobs by their parsed schedule and timezone, rather than the spelling of the cron expression. For example, */5 * * * * matches 0-59/5 * * * *, and @hourly matches 0 * * * *.

An H schedule joins a group when its resolved schedule matches. Disabled jobs and @reboot jobs are excluded. Jobs without identical schedules can still have individual run times in common; use schedule load to see those overlaps.

The endpoint

GET /schedule/duplicates

{
  "jobs": 41,
  "groups": [
    {
      "expression": "0 0 * * *",
      "description": "At 00:00, every day",
      "timezone": "UTC",
      "count": 14,
      "jobs": ["billing-export", "cleanup-tmp", "..."]
    }
  ]
}

Groups are sorted largest first. expression is the most common source spelling among the members, and description is the shared schedule in plain English. The same data backs the cron_schedule_duplicates Model Context Protocol (MCP) tool.

In the dashboards

The web dashboard's schedule load card lists the groups as clickable job chips. A chip opens that job's schedule tab. The terminal dashboard's schedule load overlay shows the top groups inline.

What to do with a group

A duplicate group is not automatically a problem: four probes that must all fire each minute are supposed to coincide. The group becomes actionable when the members are independent batch work that happened to copy the same expression. Then either spread them with H hashed schedules (one edit per job, no coordination) or give each a concrete minute from suggest a slot.