v3.14
Improved
Portals embedded in an iframe read the color scheme of the parent page and switch with it. Set a separate accent color for dark mode under Settings, then Portal.
Nothing changes for portals opened in a new tab.
Changelog
New features, improvements and fixes, newest first. Changes that break an integration are announced here 60 days ahead.
All
New
Improved
Fixed
v3.14
Improved
Portals embedded in an iframe read the color scheme of the parent page and switch with it. Set a separate accent color for dark mode under Settings, then Portal.
Nothing changes for portals opened in a new tab.
v3.13
New
The new since parameter on the replay endpoint queues every failed delivery after that time, oldest first. The portal has a matching button.
v3.12
Improved
We moved the attempt log to a new index. Searches by event ID, endpoint, status code and time range return in under a second for accounts with up to 50 million attempts.
v3.11.2
Fixed
Between August 19 and August 25, attempts that hit the 15 second timeout were stored without their request headers. Deliveries were not affected. The headers are logged again and the missing ones were restored from backups.
v3.11
New
Endpoints can list prefixes ending in .* next to exact event types. When you add a new event type under that prefix, subscribed endpoints receive it without any change on your side.
v3.10
New
Install sluice 3.10 or later and import AsyncSluice. The sync client is unchanged.
v3.9.1
Improved
When a large customer's server came back after an outage, thousands of retries arrived in the same second. Each retry delay now gets a random jitter of up to 10 percent.
v3.9.0.1
Fixed
Test messages sent from the portal were signed with a timestamp from a server whose clock had drifted. Strict verifiers rejected them. All signing now uses the same time source.
v3.9
New
After three months in beta with 140 accounts, transformations are available on every plan. Limits: 50 ms per run, 64 KB of code, no network access.
v3.8
New
Invalid events are rejected with a 422 response that names the failing field, so a bug on your side never reaches a customer. The portal shows each schema next to the event name.