October 2026

Back-to-back availability blocks are merged into one swap request

Busy days often show up in a calendar as several back-to-back blocks — a meeting with travel time either side, or the same appointment recorded in both a work and a personal calendar. Previously each of those availability entries was treated separately, so a single conflicting shift could open several automatic swap requests.

On-Call Optimizer now merges availability entries that overlap or follow on from one another, and opens just one swap request per shift for the merged block. The whole merged block is also used to decide whether a full or partial swap is requested. If you have automatic swap requests turned off, you’ll receive a single Assignment Conflict notification for the merged block instead of one per entry.

For more details see the swaps docs.

More reliable destination exports

A set of improvements to improve the robustness of schedule export to connected destinations:

• When several shorter overrides exist within a single longer On-Call Optimizer shift and only one of them is missing or wrong, just that one override is fixed. Previously all the overrides were removed and replaced by one long override. This reduces churn and API request rates.
• When an on-call product connection is unhealthy, an export now stops straight away with one clear error message instead of reporting a separate failure on each exported shift.
• Destination failure notifications are no longer held back when a destination fails for a different reason soon after an earlier, unrelated failure.
• Post-export validation checkes are now immediately retried rather than leaving the destination out of date.

For Splunk On-Call specifically:

• Changing who is on call for an existing override now updates that override in place, so cover for other schedules sharing the override is never disturbed.
• Exports for schedules sharing the same connection now run one at a time, so changes to shared overrides between schedules cannot race with each other.
• In exceptional cases when an export to one schedule cannot restore an override impacting a different schedule, the impacted schedule’s destination is immediately scheduled for synchronization.

Bugfixes and minor changes

• Fixed Splunk On-Call exports occasionally reporting a missed shift, and finishing with only partial success, when an unrelated override at the same time was removed.
• Cancelling a swap that another member has simultaneously modified now returns you to the swap and explains why it can no longer be cancelled. Accepted swaps now also appear in your Pending Requests list until they are confirmed.
• New availability blocks now check for existing swaps (including manual requests) before automatically requesting a swap to avoid duplicate swap requests for the same time period.
• Fixed Slack new shift messages failing to send for members with many shifts. Long lists now show the first 10 shifts and swaps, followed by how many more there are.
• Fixed Slack destination status messages failing to send when a destination reported several errors with long descriptive text. The message now shows the (possibly truncated) first error and links to On-Call Optimizer for the full list.
• Fixed ICS calendar feeds that don’t include a calendar name failing to load. These feeds are now listed using the calendar’s standard name, or “ICS Feed” when no name is provided.
• Fixed Slack messages beginning with an emoji (such as :link:) occasionally failing to be processed.