You do not say if this honey looking residue is a new problem with this flux or something that has just started. However you need to talk with your flux supplier.

Is this a flux that is designed to be left on the bard after processing, such as a no clean flux? The residue you describe sounds like rosin although it could be due to some kind of denaturing of a flux by an earlier cleaning process.

In either case I would be worried about long term reliablity, if this flux is still active.

Sounds like you are applying too much tacky flux during the assembly operation. Try using a Flux Transfer Plate that is 1/3 the ball diameter in depth.

This will apply a minimal and repeatable amount of flux that typically requires no cleaning.

The best option in this case would be a spot cleaning with a rework /repair defluxing agent, as the observed residue is flux. It is no surprise that the parts are failing the thermal cycling.

If the client doesn't want to send these boards through a second wash cycle, we would highly recommend using a pH-neutral solvent based cleaning agent that doesn't require rinsing and that dries fast.

Such defluxing agents come in aerosol can with brush tip or in metal canister as they are solvent based. Please contact Zestron for product recommendation and sample.

Suggestion: use nitrogen in the reflow and rework process. This will reduce the flux residue.

Typically what you are left with is the residual tacky flux that didn't evaporate off. If this is a no clean tacky flux then what is typically left over is different rosin materials. These materials are inert and no water soluble. To remove you will need to use a cleaner of some sort.

Companies that make these are Kyzen and Zestron. If the material is water soluble then the tacky flux needs to be cleaned after rework. Typically you can use water but if the water can't get under the component then again I would suggest a cleaner to help reduce the surface tension and aid the removal of the flux residue.

As for an alternative I would suggest a no clean tacky flux which leaves a clear residue. This should be cosmetically appealing and won't lead to any failures.

Rework flux not activating completely / Improper rework heating cycles can affect the robustness of the solder joint and cause thermal cycling issues

In addition, the honey-like amber flux residue needs to be examined to make sure that there is no dendrite formation and electrical shorts - a lot of rework fluxes need the right amount of heat to activate completely and do its job - because of the nature of rework, complete flux activation may not happen always and this is a potential reliability concern.

On a separate note, using rework fluxes (both tacky & liquid) that pass SIR unactivated makes sure that even if the flux were not activated, there are no electrical issues over time (due to dendrite growth).

The amber tacky residues would probably be the Rosin carrier for the Solder Paste.

I assume this is a Water Washable Paste as you say you do not want to process them through a Full wash cycle.

Functional failures could easily be the unremoved residue of the paste being conductive/Hygroscopic and I would strongly recommend cleaning them off properly in the first place back through a full wash cycle.

Partially cleaned Rosin residues will normally appear white in colour the boards may not function correctly and could easily fail in Environmental test or worst out in the field.

If the boards do not wash off correctly then investigate different cleaning mediums such as Saponifiers or branded cleaners designed for this process that assist the wetting ability of the cleaner to penetrate under the BGA. ALWAYS finish with a good DI Rinse