Product and documentation changelog

This page records material changes to GraphJSON’s product documentation, public developer experience, and customer-facing operating guidance from July 26, 2026 forward.

For endpoint compatibility, breaking-change notices, and machine-contract changes, use the API compatibility and changelog. A feature appearing here does not change the API contract unless the API changelog says so.

July 26, 2026#

Diagnostic, developer-platform, agent, workflow, and ecosystem playbooks#

Type: Documentation expansion

Added twelve advanced content areas across fifteen production guides:

• Metric-change decomposition and root-cause analysis, covering comparable windows, pipeline checks, contribution tables, mix shifts, compatible denominators, drill-down trees, and evidence
• API product and developer-platform analytics, with developer activation, endpoint breadth and depth, reliability, version migration, quota pressure, and customer value
• AI-agent, tool-use, and evaluation analytics, separating runs, steps, generations, tools, approvals, verified outcomes, versioned evaluations, cost, loops, and sensitive content
• Feature lifecycle and deprecation analytics, joining eligibility, discovery, first value, durable adoption, decline, replacement migration, removal gates, and asset retirement
• Workflow and state-machine analytics, with immutable transitions, chain validation, time in state, work in progress, rework, bottlenecks, service clocks, and mature outcomes
• Permissions, roles, and entitlement analytics, separating grants, access decisions, event-time entitlements, denial friction, utilization, stale-access candidates, and policy changes
• Referral and product-led growth-loop analytics, including sender participation, recipient activation, loop cycles, reward decisions, retention, incrementality boundaries, and abuse signals
• Mobile product analytics, covering installation identity, first observation, activation, permissions, sessions, release adoption, crash boundaries, offline delivery, and notifications
• OpenTelemetry ingestion and correlation, using a controlled adapter for selected terminal outcomes while preserving an independent full observability pipeline
• Warehouse and BI handoff patterns, with bounded aggregate extracts, raw dual writes, reverse publication, direct BI requests, immutable runs, restatement, and reconciliation
• No-code and automation-platform adapters, defining safe HTTP workflows for Zapier, Make, n8n, Pipedream, and similar tools
• four additional migration runbooks for Heap, Pendo, RudderStack, and Snowplow, covering export selection, source identity, schema mapping, time, backfill, dual delivery, and cutover

These guides explicitly preserve application, authorization, workflow, rewards, billing, observability, warehouse, and provider authorities. GraphJSON does not gain a native root-cause engine, OTLP receiver, trace viewer, agent evaluator, authorization service, workflow engine, migration exporter, warehouse connector, or managed automation-platform integration from this documentation release. No API contract changed.

Lifecycle, portfolio, planning, and enterprise delivery playbooks#

Type: Documentation expansion

Added twelve production guides that connect customer lifecycle, commercial operations, reliability, implementation, and content outcomes:

• Trial, onboarding, and customer-lifecycle analytics, with eligibility, milestones, activation, paid conversion, maturity, cancellation, churn, and winback
• Message brokers and stream-processing ingestion, covering Kafka, SQS, Kinesis, Pub/Sub, RabbitMQ, checkpoints, ordering, batching, retries, dead letters, replay, and reconciliation
• Collaboration, invitations, and team-adoption analytics, including invitation lifecycle, first collaborative value, roles, seats, multi-user activation, participation depth, and retention
• Real-user monitoring and Web Vitals, with first-party collection, navigation outcomes, LCP, INP, CLS, SPA transitions, sampling, percentiles, releases, and privacy
• Pricing, packaging, and plan-change analytics, separating eligibility, offers, paywalls, selection, checkout, billing, plans, seats, add-ons, and price experiments
• Sales pipeline and Revenue Operations analytics, covering opportunity grains, stage transitions, pipeline snapshots, qualification, handoff, velocity, maturity, and closed outcomes
• Behavioral driver and outcome analysis, with temporal contracts, entity feature rows, maturity, confounding, leakage, multiple comparisons, confirmation, and causal boundaries
• Incident lifecycle and post-incident analytics, including detection, declaration, mitigation, recovery, customer impact, communication, recurrence, and action-item follow-through
• Targets, forecasts, and plan-versus-actual reporting, with immutable plan versions, forecast vintages, scenarios, uncertainty, variance, backtesting, and partial actuals
• Multi-product and cross-product analytics, covering entitlement, product activation, attach, sequences, portfolio breadth, migration, substitution, and expansion
• JVM and .NET reference implementations, with Java, Kotlin, Spring Boot, C#, ASP.NET, typed events, HTTP clients, outboxes, bulk delivery, retries, and tests
• Content and documentation analytics, joining content catalogs, visible views, engaged reading, search, copy actions, feedback, staleness, and assisted outcomes

These guides preserve external authorities and explicitly document that GraphJSON does not provide managed message-broker connectors, lifecycle or CRM state, product entitlements, a RUM agent, incident coordination, a forecasting engine, or official language SDKs. No new product capability or API contract is implied by this release.

Cohorts, unit economics, and domain operating systems#

Type: Documentation expansion

Added twelve decision-grade guides for analytical models and product domains that were not yet covered end to end:

• Reusable behavioral cohorts and segments, with dynamic and frozen membership, event-time properties, negative behavior, maturity, snapshots, and versioning
• North Star metrics and metric trees, connecting customer value, controllable inputs, guardrails, validation, ownership, and dashboard structure
• Account analytics and customer 360, combining product, billing, CRM, support, and reliability facts at a tenant-safe account grain
• Scheduled reports and recurring delivery, covering external scheduling, Data API execution, freshness, rendering, delivery, retries, and retirement
• Sampling, approximation, and statistical uncertainty, with deterministic selection, weighting, approximate aggregations, confidence, bias, and exact-metric boundaries
• Multi-currency and FX analytics, defining transaction, settlement, functional, and presentation currencies, conversion policy, refunds, rounding, restatement, and reconciliation
• Customer acquisition and unit economics, joining governed spend, attribution, CAC, contribution margin, LTV, payback, and cohort maturity
• Product search and discovery analytics, including requests, zero results, selection, rank, reformulation, latency, assisted conversion, bots, and query privacy
• Support operations analytics, with ticket lifecycle, backlog stocks, business clocks, SLA maturity, reopen, escalation, and CSAT semantics
• Survey, NPS, and feedback analytics, covering eligibility, exposure, denominator choice, response coverage, nonresponse, follow-up, privacy, and causal limits
• Marketplace and two-sided platform analytics, with supply, demand, matching, liquidity, GMV, take rate, cancellations, refunds, and cross-side retention
• Abuse and risk-signal analytics, separating observations, decisions, challenges, outcome labels, false positives, enforcement, and privacy

These guides explicitly document the external authorities and workers required where GraphJSON does not provide a saved-cohort object, metric-tree editor, profile builder, report scheduler, FX engine, statistical engine, support system, survey platform, marketplace ledger, or risk enforcement service. No new product capability or API contract is implied by this release.

Continuous data, measurement semantics, and operational recipes#

Type: Documentation expansion

Added twelve guides for the next layer of production analytics:

• Database change data capture and continuous sync, covering outbox, CDC, polling, bootstrap, ordering, checkpoints, replay, schema drift, and source reconciliation
• Derived events and scheduled rollups, with deterministic windows, completeness watermarks, lineage, definition versions, late data, and restatement
• Sessionization and engaged-time measurement, including browser, mobile, cross-tab, identity, timeout, heartbeat, and offline-event semantics
• Where GraphJSON fits in your data architecture, defining boundaries with application databases, warehouses, observability, product analytics, and embedded BI
• Continuous reconciliation, with settled controls, independent authority, tolerances, certification states, drill-down, and failure response
• Security audit event analytics, including actor, subject, action, outcome, evidence, delivery, privacy, investigation, and authoritative-ledger boundaries
• Background jobs and queue analytics, covering logical jobs, attempts, workflows, retry cost, latency decomposition, backlog snapshots, and stuck work
• Feature flags and rollout analytics, separating evaluation, assignment, exposure, and outcomes while preserving configuration, guardrails, and rollback evidence
• Time-series baselines and anomaly detection, with explainable thresholds, comparable periods, robust deviation, missing-data states, annotations, and evaluator boundaries
• Event catalog and data dictionary, covering ownership, lifecycle, sensitivity, cardinality, source contracts, endorsement, lineage, and deprecation
• Commerce and order-lifecycle analytics, with order, line, payment, refund, fulfillment, conversion, repeat-purchase, currency, and reconciliation semantics
• Email, push, and notification analytics, separating application request, provider delivery, engagement, consent, attribution, deduplication, and channel limitations

These pages document external workers and authoritative systems where GraphJSON does not provide a managed connector, scheduler, flag evaluator, statistical model, security ledger, or messaging service. No new product capability or API contract is implied by this release.

Operational analytics, data architecture, and customer workflows#

Type: Documentation expansion

Added twelve production guides that close the next set of gaps between an initial analytics implementation and a system that can be operated, changed, recovered, and exposed safely:

• Structured application logs, including a stable log schema, severity, correlation, privacy, delivery, sampling, and investigations
• Native Data API frontend, including a server-owned report registry, tenant-safe caching, response normalization, complete React states, accessibility, and tests
• Event-time state and reference data, covering transitions, point-in-time lookups, snapshots, corrections, and current-versus-historical semantics
• Journey and path analysis, including ordered paths, noise removal, drop-off, time-to-convert, maturity windows, and funnel promotion
• Metric and SQL tests in CI, with source-controlled definitions, fixtures, exact expectations, boundary cases, and a staging smoke-test pattern
• Analytics asset change management, covering dependency inventory, change classes, parallel migration, reconciliation, cutover, and rollback
• Collection architecture at scale, including ownership, sensitivity, environment, retention, volume, tenant, and migration boundaries
• Durable raw archive and dual-write, using one canonical outbox, independently checkpointed consumers, manifests, reconciliation, and tested restore
• SLOs, error budgets, and alerts, with explicit native-alert limits, burn-rate measurements, shared-fate protection, and independent paging
• Public metrics and status dashboard, including publication contracts, sanitized snapshots, freshness states, accessible delivery, and incident fallback
• Customer health and account scoring, with interpretable components, cold-start behavior, validation, reason codes, and versioned rollout
• Customer support account timelines, including event allowlists, server authorization, correlation, pagination, redaction, and investigation auditing

The documentation index and advanced learning paths now route readers to these guides by task. No ingestion endpoint or authentication behavior changed as part of this documentation release.

Production analytics and integration playbooks#

Type: Documentation expansion

Added production guidance for the parts of an analytics system that become important after the first event and chart:

• Executable event contracts, including schemas, fixtures, compatibility rules, and CI enforcement
• High-volume ingestion and backfills, including resumability, throttling, reconciliation, and abort criteria
• Identity stitching and account lifecycle, including anonymous-to-known transitions, account switching, merges, and deletion
• Bad-data containment, correction, and replay
• Product metrics cookbook, including activity, activation, adoption, retention, resurrection, and revenue retention
• an expanded experimentation and funnel guide covering assignment, exposure, sample-ratio mismatch, power, guardrails, and rollout decisions
• File, database, and object-storage migrations
• Operating managed integrations
• production references for Next.js and Node, Python, Rails, and Go
• Accessible and localized embedded analytics

This release also added exact Data API response contracts, response examples, and validation patterns for native embedded-analytics clients.

No ingestion endpoint or authentication behavior changed as part of this documentation release.

Public developer and operational reference#

Type: Documentation and tooling

Published or substantially expanded:

• the OpenAPI and Postman toolchain
• API compatibility policy
• service behavior and operational expectations
• service status and incident communication
• security review and shared responsibility
• guides for event governance, delivery, browser/mobile collection, SQL operations, migrations, privacy, billing, and account recovery

The documentation is now served on graphjson.com/docs/... so product, API, and operational guidance share one canonical domain.

Documentation lifecycle established#

Type: Documentation policy

GraphJSON documentation now carries a review month and follows the lifecycle below. Earlier product and documentation history has not been reconstructed, so the absence of an older entry is not evidence that no change occurred.

How to read an entry#

Each entry should identify:

• Date: when the change became available or the guidance became canonical
• Type: product, API, documentation, policy, correction, deprecation, or developer tooling
• Scope: affected workflows, endpoints, roles, or deployment environments
• Action: what a customer must do, or an explicit statement that no action is required
• Links: the canonical instructions and any migration path

When a change is corrected after publication, the entry should retain the original decision and add the correction date rather than silently rewriting history.

Documentation lifecycle#

GraphJSON treats documentation as part of the product:

1. A feature or behavior change identifies its affected documentation during planning.
2. Public guidance is updated before or with customer availability.
3. Examples are checked against the current request and response contracts.
4. High-risk pages are reviewed after material API, security, billing, privacy, or operational changes.
5. A page’s last_reviewed month means it was checked for current behavior; it is not the original publication date.
6. Superseded guidance is redirected or marked with a replacement instead of being left ambiguous.

The current public documentation does not yet offer an automated changelog subscription. Teams with a business-critical integration should maintain a support contact with GraphJSON and review the API changelog before integration upgrades.

Report a gap or correction#

If the product behaves differently from the documentation, email hi@graphjson.com with:

• the affected page
• the exact step, request, or response
• approximate time and time zone
• collection or report name, when relevant
• a redacted error message or screenshot

Never send an API key, complete embed URL, payment data, or an unredacted customer payload.