firstly, there is a spectrum here from levels of resolution here in those three terms - convergence, habits, habitats. For example:
- what looks convergence at surface level is different when viewed at clade characters resolution levels ofc and what we view as convergence can again be intertwined with Pareidolia and our own human attributions.
- and the natural history and habit is so broad term that encompasses multiple axes like food-activity-hunting-reproduction-social … aka to truly find divergence in all of these at a time or some relaxed resolution levels within that constraint. and if relaxing are we relying truly on causal axes. Say if there is a social life in two distant species but one’s social life is so different from others, but functionally if the convergence happened solely because there exists social life template alone and not causal on the exact procedures involved in that social life, then should we consider these or not. (and the exact trait’s causal reasons in lot of species are very hard ofc - see my strawberry question . we observe things and we can infer why that shape is optimal or why that color has advantage but the question of “is there real causal reason why only this appeared and not all other equally working solutions and is there any pre-bias in evolution pathways and if so what?” is always a roadblock in current science and that is important when we have to deal with convergent evolution)
- lastly the habitat has resolutions like micro (rocky vs marsh) or macro(aquatic vs terrestrial), …
I think the first is already seen under relaxed levels in above examples already whether by possible mimicry of extinct ancestor or adaptive radiations or just random evolution in distantly connected populations (notice how @upupa-epops birds doesnt overlap and my snail eggs dont overlap in ranges and are different families, so they have evolved the same visual traits independently perhaps - colors or shapes ) and in regards to Pareidolia we have the superficial Apefly pupa - they are definitely not mimicking each other nor it is a functional mimicry and its not true whole Ape body in first place but so does how we cannot expect full Ape bodyplan in butterfly clade too. Such scenarios we have to end up in when you are excluding mimics and also asking for different habits/habitats explicitly. We have to let go of strong convergence there.
Now the second condition is tricky. we can arrive at partial convergence. So although we can never find a “True” bird lookalike in deep ocean (although the modern land life evolved from oceans the actual bodyplans diverged vastly in between to their environment), but in relation to piecewise convergence we can find things like Flyingfishes. The fins and wings here are equivalent for their functional purpose which again evolved from different base body plans. Another example of Sea spiders which are nothing alike spiders in some ways but that is the closest convergence that happened in such vastly different habitat. we still dont know lot of marine life so I will leave the verdict of that out there as we can always find weird complex shapes that may converge to other groups easily - say what is this marine thing even 
.
Lastly, within similar habitat but different habits, dolphin and shark are so distant (then there is extinct reptile Ichthyosaur) - but the physics places limitations eventually on overall body plans in habitat (ofc they are only similar only when viewed at higher resolutions) and then we have Australian Echidna which are different order from Porcupines where both modified hairs but have disparate foraging styles and this section in Euphorbia which comes close to Cacti, differing in microhabitats but similar functional trait evolution we have Jerboas and Australian Notomys, flying squirrel Petaurista and Australian glider Petaurus which are different orders and also different microhabitat …