Checked against Atlassian’s docs on 17 Oct 2026

Schedule Jira exports

Jira can run a rule on a clock. It cannot, on its own, leave an Excel, CSV, or PDF file of that search in History every Monday. The gap is the file, not the clock.

What a scheduled automation rule will do

Automation for Jira has a scheduled trigger. The rule can then search with JQL and act on the issues it finds: edit them, comment, transition, or send a notification. Two limits matter if the goal is “export everything open in this project”.

A Cloud rule whose JQL search passes 1,000 work items fails with an audit-log error and processes only the first 1,000. Atlassian’s write-up is A JQL search in this automation rule has exceeded the allowed maximum, updated 26 Sep 2025. The fix they give is to narrow the JQL, not to raise the export.

The automation service limits also describe a scheduled-JQL issue limit (default 1,000). If the query is larger than that, the rule editor warns you and only offers a schedule of at most four times in 24 hours. That limit is applied when you configure the rule.

The Lookup issues action is a different, smaller cap. Atlassian’s Lookup issues page sets the default at 100 issues and says not to push it past 250, because execution time grows with the list and the site can be throttled. The configuration screen on Cloud is not the same clicks as Data Center. The number is the part that matters: Lookup issues is not a bulk export.

Automation still does not write the file

There is no “export CSV” action. A rule can send JSON to a web request, which means the issue data leaves Jira for whatever URL you name. A script outside Jira can call the REST API and write a CSV. Both are real options. Both are a second system, a credential, and a place the data now lives.

The manual navigator export, including the 10,000-issue async CSV, does not have a schedule. Someone still has to open the search and choose Export. See export Jira issues to Excel for what that file actually is.

What Haulcopy’s schedule does

A Haulcopy schedule stores the JQL, the fields, the formats, the cadence, and a local time. The clock starts as the timezone on the owner’s Jira profile. Daily, weekly, or monthly. Monthly uses a day from 1 to 28 so the day exists in every month. A check every five minutes starts a schedule that is due, at most once per local day.

The search runs as the schedule’s owner, so project permissions and issue security still apply. The file appears in History for that person. Haulcopy does not email it and does not post a Jira notification. Up to 1,000 schedules on a site. You can pause one without deleting it.

Files expire after seven days. A schedule that nobody downloads is not an archive. The 17 Oct 2026 smoke on a developer site included a scheduled CSV. That is one run, not a load test of 1,000 schedules.

Set one up. The Marketplace listing is not published yet. Intended price: free up to 10 Jira users, then $1.49 per user per month, billed by Atlassian.