I watched the queue depth tick higher on the dashboard during a selective disclosure request.
I was checking a small authorized-review flow on DuskEVM this evening after the testnet settled in. Nothing complex, just a private transfer followed by the step that should surface data only for the right party. The request went through, but that disclosure piece sat visibly longer than the rest of the path.
I assumed it was ordinary testnet lag. That felt reasonable at first.
It was not that simple. Routing cleared. Hedger model availability stayed green. The payment and verification steps finished. The lag only appeared once the selective part needed to resolve.
Presence ≠ Reliability. The queue on the surface was not the real issue. Something further along held the handoff.
The path runs request to routing, model availability, payment, verification, settlement, then the disclosure layer for authorized eyes. Most of those succeeded cleanly. One did not close on the same timeline.
The dependency I keep circling is the caching decision that decides when privacy state becomes available for the next authorized call. Queue intervals and how shared infrastructure refreshes that state sit underneath and rarely get talked about.
I still do not know whether it was just residual settling after the recent launch or a tighter coupling in how the proofs propagate. I finished the Newtown tasks earlier and already ate a quiet loss on a rushed size yesterday, so I stayed with the rails instead of forcing another chart.
What holds when simultaneous authorized reviews arrive under a real onboarding spike and the economic commitment has to stay continuous?
#dusk $DUSK @Dusk
I was checking a small authorized-review flow on DuskEVM this evening after the testnet settled in. Nothing complex, just a private transfer followed by the step that should surface data only for the right party. The request went through, but that disclosure piece sat visibly longer than the rest of the path.
I assumed it was ordinary testnet lag. That felt reasonable at first.
It was not that simple. Routing cleared. Hedger model availability stayed green. The payment and verification steps finished. The lag only appeared once the selective part needed to resolve.
Presence ≠ Reliability. The queue on the surface was not the real issue. Something further along held the handoff.
The path runs request to routing, model availability, payment, verification, settlement, then the disclosure layer for authorized eyes. Most of those succeeded cleanly. One did not close on the same timeline.
The dependency I keep circling is the caching decision that decides when privacy state becomes available for the next authorized call. Queue intervals and how shared infrastructure refreshes that state sit underneath and rarely get talked about.
I still do not know whether it was just residual settling after the recent launch or a tighter coupling in how the proofs propagate. I finished the Newtown tasks earlier and already ate a quiet loss on a rushed size yesterday, so I stayed with the rails instead of forcing another chart.
What holds when simultaneous authorized reviews arrive under a real onboarding spike and the economic commitment has to stay continuous?
#dusk $DUSK @Dusk