I do not use Growatt but had the same issues, chased it for a few months and the fix was so simple & obvious and I kept going past it. Solution may be similar for you.
Note that this is applicable to my setup at the time, which was 4 Battery Packs in one Bank. 2x280AH & 2x175AH with Samlex EVO Inverter (exceedingly programmable). Generators used for this was both a 4000W Inverter Model & 9000W HD non-inverter model.
Battery Packs would charge up and as expected the 175's hit my designated 100% (3.425Vpc) before the 280's so of course their BMS' would do what they should and cutoff, balance internally etc and that was just peachy as they would level up and top off very nicely even as the amps taken dropped "for them". They would cycle on/off as balancing and toppping off occured and this is Expected in such a config.
The 280's of course take a heap more to hit Full & Top off and as the 175's dropped the amps taken and started cycling the extra immediately goes to the bigger battery packs, again this is normal & expected behaviour and all continues as it should. Then the GOTCHA that kicks the Genset appears ! It all has to do with the EndAmps / TailCurrent and transitioning to FLOAT or NOT.
3 Things are happening here:
1) BMS's reaching full start to cutoff as they should. This shoves the extra amps at whatever pack is still taking the juice, this results in surges and drops as the other BMS's flip On/Off, no genset will like that but within reason it is acceptable. ! BEWARE ! That Inverter Gennies HATE THAT ! They will cycle up & down accordingly but you have to be Very Aware of the watt / amp output limits and DO NOT push them to their spec limits ! FYI: 95% of Inverter Generators are HIGH FREQUENCY they do not take excessive abuse AT ALL ! Not even Honda !
** With the Samlex Inverter/Charger the Max Input Volts/Amps from Genset can be programmed & limited to prevent faults. IE to charge at 75A it takes 120VAC/23A from Genset, so I CAP it to no more than 25A and that still allows Passthrough.
2) Bulk Charging (Constant Current) will NOT get you to 100% SOC, at best it can get you to 95% but more likely 90% with Bulk Grade Cells. Constant Voltage Variable Current (FLOAT) get's you to 100%. THIS IS THE TRICKIEST PART ! Different gear handles the transition differently. One will use the Voltage @ Battery to determine transition to Float while another will look at Amps Taken (EndAmps) to determine the Transition. I am unsure which method is used by GW. End Amps / TailCurrent is based on the Largest Pack in the bank as it will be the last to empty & last to reach 100%. The formula is simple 200AH Battery X 0.05 = 10A. When the Packs reached 10A (collective in a bank it is different) the system should transition to FLOAT (CV). You CAN set the EndAmps to a higher value but too high and you will not have a good transition, meaning BMS' will be cutting off and cyclng.
3) The Inverter Shutdown
Oivey ! the excuses and rationalisations...
Take a genset, have it dump 120V/20A constant and then just drop the line, the genny has a crap and steps down (If Inverter Type) then all of a sudden a BMS opens and takes charge for 30 Seconds and cuts off and the next one opens does teh same & cuts off, then you have 3 or 4 BMS' doing that because the amps pushed are too high triggering Runners etc... The genset will only take that a couple of times and then trigger a fault (pending on brand that can be quick too)
-> Whenever my packs started that cycling the Inverter Genny woudl Fault and go into protect mode and not output AC till reset.
-> Whenever my Industrial Genny was hit with it, it took it & complained but kept pushing BUT the Samlex Charger would then trip & reset and THAT was FUGLY (to be extremely polite).
Some may not like / agree with what I put here but this is real world observation and testing and when I am playing with this kind of thing I can be pushing 200A+ at my bank and doing hard thrash tests. It is what it is.
The arguments over TO FLOAT or NOT TO FLOAT are also due to a lot of misunderstanding and misinformation, partly resulting from the various battery chemistries and how things are addressed for each. LFP REQUIRES CC Constant Current to get to 90-95% and then CV Constant Voltage Variable Current to finish & top off (saturate) the cells.
PS, been down those rabbit holes too often, not going there again...
Hope this helps answer some of the outstanding issues.
You will just have to determine how your Growatt Model handles the endamps/tailcurrent and transition to Float.
Good Luck.