Roadmap

Roadmap

Tracked in the open: every item is a GitHub Issue under a milestone, on the public Roadmap board.

Milestones

M0 · Foundation & DevEx

Monorepo skeleton (pnpm + Turborepo), shared contracts (zod), Prisma schema, docker-compose, CI, ADRs, repo hygiene. Goal: pnpm i && docker compose up gives a working local platform with green CI.

M1 · MVP — OTel-native Flaky Intelligence + AI RCA

The end-to-end vertical slice. OTel semantic conventions → SDK → Playwright reporter → OTLP ingestion (202 + queue) → worker → Test Identity Engine → Flaky Scoring → basic AI RCA Architecture → tRPC query API → dashboard (runs, test detail, flaky board, RCA) → GitHub Action + PR bot → dashboard auth + onboarding → e2e demo. Goal: a 60-second demo that finds and explains flaky tests without blocking CI.

M2 · Deep Observability

Full OTLP trace ingestion (test → step → network → browser), artifact pipeline (screenshots/video/HAR/trace), trace waterfall viewer, suite-health dashboard, trend analytics via rollups, notifications (Slack/Discord/email), PR quality gate (base-branch comparison), shard/parallel correlation, ownership mapping.

M3 · Test Intelligence

Failure-signature clustering, known-issue fast path (no-LLM), retrieval-augmented RCA, auto-quarantine workflow, advanced test identity (rename + parameterized), anomaly/regression alerting, flaky MTTR + DORA-style metrics, RCA eval loop, historical backfill/import.

M4 · Platform & Extensibility

Multi-framework reporters (Jest/Vitest, pytest/JUnit), plugin architecture (ingestion + analyzer hooks), public REST API + webhooks + OpenAPI, full flakemetry CLI, more CI integrations (GitLab/CircleCI/Jenkins), data export + retention/deletion.

M5 · SaaS & Scale

Multi-tenant hardening + isolation tests, RBAC/SSO/SAML, columnar span store (ClickHouse/Timescale), usage metering + open-core boundary, cloud deploy (Helm/Terraform), security hardening + threat model, load/soak testing + SLOs, backup/DR.

M6 · Community, Docs & Launch

Public docs site, landing page + branding, OSS launch (Show HN / Product Hunt / dev.to) + blog series, contributor program (good-first-issues, Discussions).

Sequencing rationale

• M0 → M1 is the critical path. Everything else builds on a trustworthy flaky signal.
• M2 depth feeds M3 intelligence — better trace context ⇒ better RCA and clustering.
• M4 breadth (more frameworks/CIs) comes after the core is proven, to avoid spreading thin early.
• M5 hardening is deferred until there's something worth scaling.
• M6 launch ships once the product tells a complete story.

How to follow along

• Issues — filter by milestone: or area: / type: labels.
• Roadmap board — Kanban across milestones.
• Discussions — ideas and design debate.