Ecosystem health metrics

(Section: Ecosystem and Network)

1) What this article is about (summary)

Ecosystem health is a collection of indicators that reflects sustainability, reliability, liquidity, interoperability, security, economics and engagement of network participants (operators, providers, studios, affiliates, nodes/chains, communities). Below is the system framework: measurement levels, a list of KPIs with formulas, the EHI composite index, target thresholds (SLO), alert rules, dashboard templates and a practical playbook of reactions.

2) Measurement level map

1. Infrastructure and network: availability, latency, throughput, errors.
2. Protocol/interoperability: success of cross-chain/cross-service operations, version compatibility, proportion of compatible nodes.
3. Product and users: activity, retention, conversion, traffic quality.
4. Economics and liquidity: turnover, liquidity depth, spreads/commissions, payment delays.
5. Community and partners: input from developers/studios, NPS, partner onboarding pace, quality of integrations.
6. Compliance, risk and safety: incidents, fraud rates, KYC/AML passage, sanctions/geo-risks.

3) Basic KPIs (with short formulas)

3. 1 Infrastructure and network

Uptime of services (%) = 100 × (Running time/Total observation time).
p95/p99 Latency (ms) - by key API/gateways/nodeendpoints.
Error Rate (%) = 100 × (5xx + explicitly fatal 4xx )/All requests.
Saturation: CPU/RAM/IO/quotas - time share> 80%.
Backpressure Events: number/day.

3. 2 Protocol and interoperability

Cross-Chain/Inter-Service Success (%) = 100 × Inter-Chain/Inter-Service Success/All Attempts.
Median Finality - to irreversibility/confirmation.
Version Compatibility (%) - Percentage of nodes/SDKs on supported versions.
Rollback/Reorg Rate - frequency of rollbacks/conflicts.

3. 3 Product and users

DAU/WAU/MAU (rationed by cohort/region).
Retention D1/D7/D30 (%) - cohort.
Activation Rate (%) = Activated/New.
Conversion Funnel: Visit→Reg→KYC→1st Action→Repeat.
Quality of Traffic (QoT): the proportion of valid traffic after anti-fraud.
Session Success (%) - the proportion of sessions without critical errors.

3. 4 Economics and liquidity

GTV/Volume - gross volume of operations.
Liquidity Depth - the median amount of available liquidity in peak hours.
Payout SLA Hit Rate (%) - share of payments ≤ target time.
Cost-to-Serve (CTS) = Operating Cost/Number of Successful Operations.
Take Rate (%) - commission/margin on volume.
Dispute Rate (%) - Disputed/Disputed Transactions.

3. 5 Community & Partners

Partner Activation Velocity - New Integrations/Week.
SDK/Plugin Adoption - installations, upgrades/version.
Community NPS/eNPS - Quarterly.
Contribution Index - pool prompts/releases/add-ons from third-party teams.
Docs Health - fullness, freshness, time to answer a question in the community.

3. 6 Compliance, risk and safety

KYC/AML Pass Rate (%) - the share of those who passed the deadline.
Fraud Rate (%) - confirmed fraud/all transactions.
Incident Rate - SEV, MTTR/MTTD levels.
Policy Coverage (%) - the proportion of threads with active DLP/PII controls.
Geo/Regulatory Coverage - markets where local requirements are met.

4) Composite Health Index: EHI (Ecosystem Health Index)

The idea: a single score of 0-100 for stakeholders.

1. Normalization: bring all KPIs to the scale [0... 100]:

Min-Max with percentile truncations (e.g. P5-P95), or

Z-score → CDF → [0…100].

2. Weight model (example):

Infrastructure - 25%

Protocol/Interoperability - 15%

Product/User - 25%

Economy/liquidity - 15%

Community/Partners - 10%

Compliance/safety - 10%

3. Formula:

'EHI = Σ (Block _ Weight × Average (Normalized Block KPIs)) '

4. Interpretation scale:

85-100: "Excellent" (growth with risk margin)

70-84: "Stable" (controlled risks)

55-69: "Vulnerable" (need point improvements)

5) Leading and lagging indicators

Leading: QoT, Activation Rate, time to finalization, CTS, share of nodes on the new version, Docs Health.
Lagging behind: MAU, GTV, Take Rate, NPS, Dispute/Fraud Rate.
Balance the portfolio: 60% leading, 40% lagging for preventiveness.

6) Thresholds (SLO) and alerting

• Uptime ≥ 99. 95 %/30d; p99 latency ≤ 400 ms; Error Rate ≤ 0. 2%.
• Cross-chain success ≥ 99. 5%; Median finality ≤ 6 с.
• Payout SLA hit ≥ 98%; Dispute ≤ 0. 3%; Fraud ≤ 0. 1%.
• KYC ≤ 10 min in 95% of users.
• Docs updated ≤ 14 days from release; median first response in the community ≤ 2 hours.

• SLO Burn Rate 1-hour> 14 × - Pager; 6-hour> 6 × - Pager; daily> 3 × - ticket + analysis.
• Always specify owner, deadline and "done" criteria.

7) Segmentation and sections

By country/jurisdiction, partner types (operators, studios, affiliates), infrastructure clusters, SDK/node versions, traffic channels, product types (slots/live/sports/financial transactions), devices.
Cross-sectional filters and cohort-to-cohort comparisons are mandatory for each metric.

8) Dashboards (layouts)

A. Daily Ops (real time/hourly)

Uptime, p99 latency, Error Rate, Cross-chain success, Incident SEV, Payout SLA, Fraud spikes.
Service card (green/yellow/red), payment/verification queue.

B. Weekly Product/Partner

Activation/Retention, QoT, конверсия KYC→1st Action, Partner Activation Velocity, SDK adoption, Docs Health.
Channel mix and LTV early (proxy).

C. Monthly Strategy

MAU/WAU, GTV, Take Rate, CTS, Dispute/Fraud, NPS, Contribution Index, Geo Coverage, EHI динамика.
Ladder of risks and "traffic light" for each block.

9) Data sources and quality

Telemetry: logs/metrics/trails, product events (event bus), nodes/validators, payment and partner APIs, KYC/AML providers, Service Desk/incidents, NPS/DevRel surveys.
Data Quality KPI: completeness, freshness (lag), uniqueness, consistency of schemes, share of "uncertain" statuses. Enter a separate DQ Score and do not mix it with EHI.

10) Anti-metrics (vanity & traps)

DAU without cohorts/regions; "average conversion" without channels; GTV without returns/disputes; uptime without taking into account critical endpoints; "number of integrations" without productive activity; "number of commits" instead of the value of releases.

11) Playbook reactions (cheat sheet)

• Enable degradation modes (read-only, caching, limits), expand horizontally, enable queue prioritization; post-mortem at 24 o'clock

• Check versions, fee/limits, retrays with idempotency, versioning of schemes; roll hotfixes, upgrade nodes/SDK.

• Path analysis KYC→1st Action, "time-to-value," friction; A/B tests of onboarding, content/localization, packaging of offers.

• Redistribute pools, add providers, automate revaluation, enable predictive calculation of cash gaps.

• Tighten scoring, limits/velocity-check, manual high-risk review, model training on fresh patterns.

• DevRel programs, grants/bounties, SDK/dock improvements, monthly office-hours, support acceleration.

12) Goal templates (OKR, example per quarter)

KR1 (Infra): p99 latency API ≤ 350 ms; uptime ≥ 99. 97%; Error Rate ≤ 0. 15%.
KR2: Cross-chain success ≥ 99. 7%; median finality ≤ 5 s; ≥ 80% nodes on LTS.
KR3 (Product): D7 retention + 3 pp; Activation + 5 pp; QoT + 4 p.p.
KR4 (Economics): Payout SLA hit ≥ 99%; CTS −10%; Dispute ≤ 0. 25%.
KR5 (Community/partners): + 15 active integrations; Docs Health 90/100; NPS ≥ 45.
KR6: Fraud ≤ 0. 08%; MTTR ≤ 30 min (SEV-1); 100% of critical streams are covered by DLP/PII.

13) Implementation in data (reference pieces)

Pseudo-SQL Active Users by Region

sql
SELECT date, region, COUNT(DISTINCT user_id) AS dau
FROM analytics. events
WHERE action IN ('session_start','game_start','bet_place','deposit')
AND date BETWEEN:from AND:to
GROUP BY 1,2;

Cross-chain success

sql
SELECT date_trunc('hour', ts) AS h,
100. 0 SUM(CASE WHEN status='success' THEN 1 END)/COUNT() AS success_pct
FROM interop. tx
WHERE ts >= now() - interval '7 days'
GROUP BY 1;

Payout SLA

sql
SELECT date::date,
100. 0 AVG(CASE WHEN payout_sec <= target_sec THEN 1 ELSE 0 END) AS sla_hit
FROM payouts. metrics
GROUP BY 1;

Preparation for EHI (min-max)

sql
SELECT kpi, 100. 0(value - min_v)/(max_v - min_v) AS score_0_100
FROM kpi_current
JOIN kpi_ref ON kpi_current. kpi = kpi_ref. kpi;

14) Glossary

EHI - Ecosystem Health Integral Score 0-100.
SLO/SLA - target quality levels/contractual levels.
Finality - time to irreversibility/confirmation of the transaction.
QoT is a traffic/user source quality metric.
CTS - unit cost of service.
Burn Rate (SLO) - the rate of "burning" the error budget relative to SLO.

15) Implementation checklist

1. Record KPIs, sources, owners, frequency.
2. Define the SLO and alert thresholds (1h/6h/day).
3. Set up dashboards: Ops (day), Product (week), Strategy (month).
4. Implement the EHI and publish it as per regulation (e.g. weekly).
5. Review metrics, weights and SLOs quarterly.

Bottom line: This framework provides common ground for infrastructure, product, partner and compliance teams, reduces "blind spots" and allows signals to be turned into quick, concerted action - before ecosystem weaknesses become a problem.