privacy · proofs · compute
v3.2 · checksummed

Appendix B. Practical KPIs & Telemetry Templates

v3.1
Download PDF Table of Contents Cite / Checksums
Cite this section

Copy/paste (plain text): 

Jason St George. "Appendix B. Practical KPIs & Telemetry Templates" in Next Generation Stores of Value: Privacy, Proofs, Compute. Version v3.1. /v/3.1/read/appendix/b-telemetry-templates/

Practical KPIs & Telemetry Templates

This appendix collects the thesis’s named instruments into one table — what each measures, in what unit, from what data, how often, on which public board it surfaces, which red line reads it, and at what threshold — and then lists the operational checklists by domain. The table is the reconciliation: a metric that appears in the body but not here, or here with a different definition, is a defect to report. Definitions are drawn from the body and the formal appendices (Appendix A: Formal Model of Verification Asymmetry & VerifyPrice, Appendix H: Formal Model of Market Realization, Wrapper Flows, and Price Capture); where a threshold is a red-line trigger it is the red line’s own figure, and where the body publishes no number the cell says so rather than inventing one.

Instrument Table

Operational Checklists by Domain

The table above holds the named instruments. Operators also need the ordinary telemetry beneath them, grouped here by domain. None of these carries a red line by itself; each feeds one of the instruments above or one of the twelve boards (§23: Extended Telemetry).

Prover markets:

Queue depth, clearing price per proof type, cancellation rates, p50/p95 VerifyPrice.

Compute consensus pilots:

MatMul throughput (GFLOPs/s), verifier cost ratio r(W)r(W) with its modality band, invalid-work rejection rate under adversarial load, adversarial acceptance-suite pass rate (§19: Layer 4: Truth & Work).

Privacy settlement:

Atomic swap success %, time-to-finality, stuck-flow causes by fmixf_{\text{mix}} category, wallet UX friction metrics.

Decentralization & capture risk:

  • Nakamoto coeff. (by stake/hash/prover share); top-N concentration; Gini.

  • Entry latency: time-to-first-proof/mined-block for a new node.

  • Geographic/ASN diversity; % over Tor/VPN; relay/MEV-builder diversity.

Liveness & safety:

  • Reorg/orphan rate; effective finality time (p50/p95); incident MTTR/MTBF.

  • SLA attainment (% proofs/settlements meeting latency/uptime targets).

  • DoS/churn resilience: success % under adversarial load mixes.

Verify/produce economics:

  • Verifier cost ratio (verify/produce) by workload; energy per proof/inference.

  • Hardware mix & utilization (GPU/CPU/ASIC) vs. throughput; cost/FLOP ($/GFLOP-verified).

  • Failure taxonomy: prover faults, invalid proofs, circuit timeouts.

Privacy rails health:

  • Anonymity-set size & churn (shielded pool/epoch); linkability regressions found.

  • Swap slippage & fees (p50/p95); cross-venue route diversity; failure-mix root causes, classified under the fixed fmixf_{\text{mix}} taxonomy of Appendix A: Formal Model of Verification Asymmetry & VerifyPrice, Definition 5, rather than free text.

  • UX friction: successful first-run rate, steps/clicks to complete, abandonment rate.

Bridges & cross-chain settlement:

  • Atomicity violations (0-conf leaks, partial completes); retry rate; liquidity depth per route.

  • Watchtower/guardian coverage; light-client verification share vs. multisig.

Provenance & policy demand:

  • C2PA [C2PA spec] attach rate, strip rate by platform; “verified views” share.

  • Proof volume indexed to policy milestones (EU AI Act [EU 2024] dates); % budgeted vs. ad-hoc spend.

Security & abuse:

  • Sybil detection hit rate; staking/miner collusion alerts; MEV/censorship events.

  • Key/attestation hygiene: validator/prover key rotations, slashing/penalties.

Token & issuance telemetry:

  • Realized issuance vs. schedule; FeeCoverage (retained fees, burns excluded) of security budget; % rewards tied to useful work.

  • Velocity vs. locked/collateralized supply; SoV signals (holder age bands, exchange outflows).

Customer success:

  • Enterprise proof spend: $/month, SLA breaches, churn; Net Proof Retention (NPR) — for the cohort of customers that were paying for proofs in a base period, their proof spend in the current period divided by their base-period spend, computed on receipt-tagged budgets and excluding new customers; NPR >1> 1 means the existing base grew its verification demand net of churn and down-sell, the proof-market analogue of net revenue retention.

  • Time-to-integration (SDK→\toprod); support tickets per 1k proofs/settlements.

Hardware honesty:

  • Share of receipts tied to open or sampled hardware profiles (by workload and volume).

  • Lot-sampling coverage: percentage of lots and devices subject to destructive audits (SEM/optical imaging, side-channel tests), and maximum time since the last audit per profile.

  • Incidents and deprecations: count and severity of profile-level compromises, time-to-deprecation, and migration progress away from affected profiles.

New here? Start with the one-minute version.

Tip: hover a heading to reveal its permalink symbol for copying.