Is it normal to write automation for compression/limiter?Or you can automate everything if it's right.
As for your comment
i want to automate master limiter but not sure 100%
I'd actually say this is probably a bad idea. The purpose of a master limiter is, well, to limit the master level. It isn't really an effect so much as a safeguard. IMO the master limiter should be set up in such a way that it makes best use of the distribution mediums dynamic range. There are reasons why this is traditionally done by the mastering engineer, not by the mixing engineer.
If you find that the limiter is pumping excessively in one section but hardly does anything in another, then this is something that might make sense to tackle with automation – but not of the limiter but of the mixing levels. Obviously, the limiter only takes action when the signal actually gets to the limit. This may be down to a single track in your mix, in which case it can make sense to automate the level of that track down at this spot. Or maybe you should work with a multiband compressor before the limiter. Or...
is it normal to write automation for compression
yes it is, in particular for channel compressors. In fact I tend to prefer automating compressor thresholds over automating channel levels, because it better preserves the perceived loudness of a channel while still adjusting the “firepower” of that channel. In particular, to get back at the master-limiter example: if one channel pumps the master limiter at one place in the song, it's probably a good idea to tame that channel with an individual compressor. And automating the threshold of that compressor can be a good way of adjusting how much the channel drives the master compression, without incuring a very notable change in the mixing ratios.