A small interactive map that visualizes meaning and schema drift across distributed systems over time.
This project is part of my Systems of Trust Series, exploring how distributed systems maintain coherence, truth, and alignment across services, schemas, and teams.
The goal of this map is to make truth drift legible — not just event validity, but meaning: how shared concepts change, fork, or degrade as they pass through pipelines, teams, and services.
Distributed systems don’t usually break because a single service fails.
They fracture because shared meaning drifts.
Over time, small changes accumulate:
- fields renamed or repurposed
- enums forked between services
- optional fields that are “required in practice”
- different teams interpreting the same event differently
These shifts create:
- conflicting views of “what happened”
- silent schema mismatches
- harder incident analysis and debugging
- inconsistent analytics and decisions
This map visualizes how those drifts emerge across services and time.
The prototype includes:
- Semantic Drift Timeline – see how a field’s meaning changes across versions and services
- Schema Divergence Map – track added, removed, or repurposed fields over time
- Service Alignment Graph – highlight which services still agree on a concept and which have drifted
- Invariant Mismatch Flags – surface different rules applied to the same field or event
- Downstream Impact Summary – rough Low / Medium / High impact for analytics, ML, and incident analysis
- Lightweight client-side experience – static HTML + JS, no backend required
This tool is intentionally minimal and aimed at conceptual clarity, not a full metadata catalog or governance platform.
[Source of Truth]
(canonical event + meaning)
|
(rename, repurpose,
add/remove fields)
|
v
[Service A Interpretation]
|
(local invariants,
partial upgrades)
|
v
[Service B Interpretation]
|
(schema forks, enum drift)
|
v
[Service C Interpretation]
|
v
Downstream Consumers
(analytics, ML, dashboards,
incident review, audits)
Even if services are “up” and requests succeed, drift in meaning can quietly erode trust:
- teams disagree on what a field actually represents
- dashboards and ML models are trained on incompatible definitions
- different services enforce different invariants on the same concept
- incident analysis becomes a negotiation of interpretations, not facts
- governance docs lag behind reality
This tool focuses on the semantic layer of trust: not just “is the event valid,” but “do we still agree on what it means?”
Each element of the map corresponds to a real architectural concern:
The same field (or event name) comes to represent slightly different concepts between services.
Example: status meaning “lifecycle state” in one service and “billing state” in another.
Schemas change version-by-version, but not all services upgrade together. This leads to:
- old services emitting legacy fields
- new services reading or writing fields differently
- “optional” fields that become required in downstream assumptions
Two services can share a field name but diverge on:
- type (
"123"vs123) - enum values (
"ACTIVE"vs"active") - cardinality (single value vs list)
These mismatches show up as fragmented logs, broken joins, and inconsistent analytics.
When meaning diverges, systems built on top of that data — attribution, billing, ML, customer journeys, audits — all quietly inherit the drift.
The Truth Drift Map makes these conceptual differences visible instead of implicit.
Main repo:
https://github.com/rtfenter/Systems-of-Trust-Series
MVP implemented and active.
The prototype of this map will focus on core mechanics needed to demonstrate truth drift at the schema/meaning layer, not a full production metadata catalog.
Everything runs client-side.
To run locally (once the prototype is implemented):
- Clone the repo
- Open
index.htmlin your browser
That’s it — static HTML + JS, no backend required.