There are impacts/considerations from a fault tolerance/recoverability and managability perspective:
"You cannot remove a single disk from a disk group. You must remove the entire disk group to make modifications."
"A single disk failure causes the entire disk group to fail."
Additional compute overhead of up to ~5%.
"The processes of deduplication and compression on any storage platform incur overhead and potentially impact performance in terms of latency and maximum IOPS."
As to the extent of possible performance hit, I cannot find much similar to compare to this online but this blog would suggest it may be significant if using RAID1 as the FTM:
However as pointed out in the comments of the above by author and elsewhere (spiceworks) the jury seems to be out on whether these forms of synthetic test are telling the full picture - additionally I workload type and/or IO size can be significant factors and I see no reference to what was used in the above tests.
Some other good examples of things that may need to be considered can be found here:
So yes, there are a few potential costs to pay for using dedupe+compression but in my opinion (based on working with ~100 clusters/customers per quarter) these are reasonable trade-offs - unless something specific is required that conflicts with any of the above points.