Sibilance can be sourced in several factors... The first is the vocalist - some people are more predicated to it than others, there are also some mics that have inherent character in the frequency range where sibilance occurs - and that range isn't necessarily always one with a narrow Q either; I've found fundamental sibilant peaks occuring anywhere between 3k and 7k ... and that's a fairly broad range.
Most traditional de-essers were a form of side-chain gain reduction - in its most basic form, a compressor with its detection circuit being triggered by a specific frequency range.
The downside to traditional de-essers is that they also tended to attenuate high frequencies that weren't at the source of the problem, so often they would take good things away, too, such as the "sparkle" and "silk" of a vocal track.
So far, the best way I've found to reduce sibilance is to actually go into the wave form itself, and to manually edit sibilance - either bringing down the volume, or by fading in from a section of the waveform which always identifies it...it has a very distinct look to it - it looks like a football -(an American football, I mean, LOL), sort of oval looking, with a "even" symmetrical spacing of its inner waves... After you identify it once, your eyes will seek it out and recognize it easily afterwards:
https://jslhr.pubs.asha.org/data/Journals/JSLHR/929530/1260fig1.jpeg
At first, manually editing it can seem tedious, but you get to a point where you get an entire track manually de-essed in about five minutes, no more than ten for longer tracks.
This isn't a to say you can't use plugs designed for this purpose; newer versions use a form of spectral editing... And this type of editing range is used to attenuate all sorts of problematic frequency based issues, such as " squeaks" that happen when fingers move over strings on acoustic guitars, clicks, pops, and other mouth noises, even crackle and static from vinyl records.
Most are available for full functioning trials. The best "auto" de-esser I've heard yet us from Fabrice Gabriel, who makes the Eiosis De-Esser, ( Fabrice does a lot of plug in modeling for Slate Digital).
You can try switching up mics at the source, but if the vocalist happens to have heavier sibilance naturally, that's not an easy thing to correct, because it can be caused by things like air gaps in between the teeth, or even the way someone just naturally speaks or sings. You might try backing them off the mic a bit, but then you're chancing a tonal change somewhere else in the spectrum which may - or may not sound good.
Thus far, manual editing has worked the best for me.
YMMV, though. Ñ;)
D.

Since a recent project demanded that I manually edit practically every line in sixteen songs it has become pretty routine for me to do it on most other vocal tracks. Once you get into it your pace picks up considerably. It can get a little tedious. I do this vocal editing in the mix, so I'm actually balancing the sibilance to the rest of the mix.

Pro Tools (among others) has a great feature for this, Clip Gain Line editing. It's like any normal automation line but it affects the audio clip upstream of the channel processing like inserts, pan and volume, and it modifies the waveform display accordingly.

Use spectral editing in REAPER, or even Audacity if you need to do it manually. Great for denoising and de-essing.

besides the other good advice, i would add to have the singer try to ease the sibilance as much as possible during tracking. theres an old trick where you tape a pencil to the mic so it deflects a little bit of the sss blast. also the singer can just not sing there ess's directly into the mic, just momentarily sing a little off axis. if your brave you can manually ride the gain or eq during tracking, but theres more sophisticated approaches as have been mentioned. its easy enough to punch in problem words, while tracking, which is arguably faster and with less artifact than editing. plus the singer does the work, the engineer editing. other benefits include a nice finished take/comp immediately when the tracking is done, and the singers will improve their techniques, making things cheaper for them, and better overall.