FlashDuty supports pushing real-time incident and alert changes to third-party systems, such as ticket platforms or incident auto-healing systems, enabling information synchronization and accelerating incident resolution.
Currently, only HTTP/HTTPS protocol with POST method is supported, and the request payload must be in JSON format.For example, if your service URL is https://example.com/webhook?a=a, the actual push request would be:
You can specify to subscribe to events from a specific channel, and filter event types. Note that incident and alert event types are different, please refer to the corresponding documentation for details.
The service must respond within 1 second, otherwise it will be considered as a failed response.
2.
Will pushes continue after a failed push?
Currently, FlashDuty pushes at most once. A retry mechanism may be introduced in the future, and retries may also occur due to middleware timeouts. You need to handle idempotency properly.
3.
How is push order guaranteed?
In theory, events for the same alert are pushed in chronological order, but retries and other situations may cause out-of-order delivery;
The service can filter based on event_time. If a later event has been received, earlier events can be filtered out. Each push carries the latest and complete information, occasional event loss is tolerable.
