I've seen this delay as well, in the same situation.
It only happens when I use a Pico to control a Caseta dimmer via Hubitat. In that case there is a delay of 1~2 seconds between the button press and the response from the dimmer. For all other cases, the response is almost instantaneous -- usually less than 1/5th of a second (which is what you would expect from local processing).
For example:
- Lutron Pico --> Hubitat --> Z-wave dimmers: almost instant response
- Z-wave button --> Hubitat --> Lutron dimmers: almost instant response
- Lutron Pico --> Hubitat --> Lutron dimmers: 1~2 second delay
- Lutron Pico --> Lutron dimmers: instant response (naturally) , but with a loss of fine-grained control (pushed vs. held, turn on to last dim level, etc.)
The delay does not show up in the Hubitat logs. It also does not show up in the Lutron logs (which you can snoop via a separate telnet session with timestamps enabled).
So it appears that the delay is internal to the Caseta bridge, and only occurs when there is a ~DEVICE statement (e.g. the Pico button press) immediately followed by an #OUTPUT statement (e.g. Hubitat telling the dimmer what to do).
I suppose if you had two Caseta bridges, you could get rid of the delay by putting all the Picos on one and all the switches / dimmers on the other.
