update(Kinesis): Add health check interval jitter config

[Meggielqk]

emqx/emqx#15387

[Meggielqk]

The default value for the health check interval jitter config is 15 second for Kinesis, while it is 0 millisecond for other Sinks on Dashboard. Are there any reason for the difference? @thalesmg

[thalesmg]

The default value for the health check interval jitter config is 15 second for Kinesis, while it is 0 millisecond for other Sinks on Dashboard. Are there any reason for the difference? @thalesmg

It's basically this comment, plus the behavior described in the changelog that introduced it:

If they receive a throttling response (e.g.: LimitExceededException), they will also maintain their current status.

In other words, if there's throttling from Kinesis and some actions are in the disconnected state, and if these disconnected actions keep firing their health checks after others under the same connector, they will not change their status. Jitter adds a random delay so that, eventually, these disconnected actions have the chance of running before others and hopefully switch to connected.

[Meggielqk]

chance

Thank you for the answer. Could you please verify if the description for this config is correct?

[thalesmg]

Thank you for the answer. Could you please verify if the description for this config is correct?

This comments seems to still be in the field description:

This helps avoid burst traffic, reduces the risk of triggering downstream rate limits (e.g., API throttling), and improves system stability.