We list every fault that a customer could see, what it did, and what we changed.
Incidents that affect customers are listed here within five business days of the fix. Newest first.
5 entries. Latest: 11 November 2026.
1 to 11 November 2026
Entries shown as undated
Fixed
What happened
After a change to date handling, an entry could be shown as Undated although its page printed a date.
Effect
At most 2 entries on one customer matter.
Fix, 11 November
Corrected for new readings. The 2 stored entries stay labelled Undated. The full test suite now blocks every release.
10 November 2026
Two stalls, unsaved changes
Fixed
What happened
The service stalled twice, for about four minutes each (15:11 to 15:15 and 16:05 to 16:09 AEST), while one very large chronology was rebuilt after each edit. The watchdog restarted the service both times. Changes saved during the stalls failed without telling the user.
Effect
One customer. 195 save attempts failed and 123 entries were left unsaved. No stored data was lost. We told the customer on 11 November and gave them a list.
Fix, 10 to 11 November
A failed save is now shown and kept on the page; an edit no longer reloads the chronology; the chronology is built off the main service thread.
9 to 10 November 2026
Slow downloads
Fixed
What happened
PDF downloads took 49 to 97 seconds.
Effect
Downloads were slow. No data was affected.
Fix, 10 to 11 November
An unchanged matter now downloads in under 3 seconds.
7 November 2026
Three faults on one day
Fix due before 1 December
What happened
Some scanned documents with very large pages were refused; the service stalled for about two minutes and restarted by itself; one brief was read four times instead of once, which replaced its entries while the examiner was working.
Effect
One customer. No data was lost.
Fix
Very large pages are now read at reduced size and listed as such (8 November). A change that guarantees a single reading is due before 1 December; until then every customer reading is watched.
5 November 2026
Service stopped for about 36 minutes
Fixed
What happened
The files of one brief arrived over a minute and a half. Reading started before the upload was complete, and a “chronology ready” email went out when 8 of 27 files had been read. The second reading stalled the service, and nothing restarted it.
Effect
One customer. No data was lost; every file was read after the restart.
Fix, 5 to 6 November
Reading now waits 5 minutes after the last upload, or until the customer presses “Upload complete: build chronology”. An email reports files added later. A watchdog restarts a stalled service.
