bounoable
commented
2 years ago Defaults
Heuristics completely pulled out of thin air: fixed cap of 5s if the debounce timeout is <= 2.5s, for higher timeouts simply double the timeout.
- 5 seconds should be a reasonable default for most applications
- doubling the timeout should effectively prevent the debounce from triggering more than twice
| Debounce | Default Cap |
|---|---|
| none | none |
| 100ms | 5s |
| 200ms | 5s |
| 500ms | 5s |
| 1s | 5s |
| 2s | 5s |
| 2.5s | 5s |
| 3s | 6s |
| 4s | 8s |
| 5s | 10s |
| 10s | 20s |
| 30s | 1m |
| 1m | 2m |
| 1.5m | 3m |
| 2m | 4m |
| 5m | 10m |
Problem
Currently, the
Debounce(time.Duration)option of the continuous schedule has no cap on the total wait time before a projection job is triggered. Application that continuously publish a lot of events could cause infinite buffering of projection events.Proposal
Add a
DebounceCap(time.Duration)option that specifies the wait cap. Use a sensible duration as the default (5s maybe?).