[Feedback collector] Identify if dealbot should do more direct calls to Synapse rather than through Filecoin Pin

[BigLep]

Background

As part of resolving #411 (comment), it was raised that we had extra steps to address a Deablot issue because we had to go synapse-core + synapse-sdk → filecoiin-pin → dealbot rather than synapse-core → dealbot. We didn't make any changes then, but this issue was opened to evaluate to collect feedback/anecdotes on whether we have a conviction that we'd come ahead from a cost/benefit regard (accounting for code changes and ongoing maintenance) if we made this switch.

[BigLep]

@hugomrdias : I created this issue after the 2026-03-31 conversation. Feel free to add anecdotes as you see them.

I'm good to evaluate when we have more observations.

My take at the moment is that it's not ideal to be slowed down to respond to issue because of the extra hop through filecoin-pin. I do think it has the upside of being a forcing function to keep filecoin-pin updated.

[SgtPooki]

one anecdote spanning multiple issues and upstream prs:

• feat(warm-storage): add hasActivePieces and use it in resolveByProviderId synapse-sdk#852
• perf(storage): bound resolveByProviderId fan-out with an ordered sliding window synapse-sdk#853
• Calibration eRPC 429 storms: deal flow resolveByProviderId fans out O(active data sets) eth_calls per deal #606

however, i think we should close this.. the dogfooding is a forcing function for improving the system as a whole