Flipt
Flipt is an open-source, self-hosted feature-flag management platform designed for engineering teams that want control over rollout, evaluation performance, and Git-driven workflows. It provides boolean and multi-variant flags, targeting rules, percentage rollouts, and supports both server-side and client-local evaluation via REST and gRPC APIs.
Flipt is aimed at teams that operate their own infrastructure (especially Kubernetes), need low-latency evaluations for high-throughput services, or prefer keeping feature-flag data and change history on-premises. It can be run lightweight for evaluation or production-grade with PostgreSQL, observability hooks, and an official Helm chart.
Use Cases
- Engineering/platform teams running services in Kubernetes who want an internal feature-flag service with repeatable deployments and observability.
- SREs who need predictable, performant flag evaluation for high-request-rate backends using gRPC or REST.
- Teams that want to treat flags as code: store flags in Git, review changes, and keep a single source of truth alongside application code.
- Product and QA teams using flags for progressive rollouts, targeted experiments, and safe canarying without redeploying code.
- Organizations with compliance or data residency requirements that mandate self-hosting rather than a SaaS provider.
Strengths
- Open-source and self-hostable: Run on your infrastructure to avoid mandatory cloud lock-in and retain control over data and upgrade cadence.
- Git-native workflows: Flag-as-code approach enables versioning, reviews, and traceability when integrated with Git.
- Performance-focused APIs: gRPC and REST evaluation endpoints and options for local evaluation suit low-latency and high-throughput systems.
- Flexible flag model: Boolean and multi-variant flags with targeting and percentage rollouts cover typical rollout and experiment needs.
- Production deployment path: Official Helm chart and Kubernetes guides make it feasible to deploy repeatably in K8s clusters.
- Pluggable storage: In-memory for testing, PostgreSQL for production durability and backups.
- Observability: Prometheus metrics and health endpoints integrate with existing monitoring and alerting workflows.
- Commercial upgrade path: Paid tiers add RBAC, audit logs, enhanced Git workflows and vendor support for teams that need enterprise controls.
Limitations
- Kubernetes-first complexity: The recommended production path favors Helm and K8s—teams without K8s experience will face an operational learning curve.
- Docs gaps for advanced cases: Documentation can be sparse on complex setups (advanced Postgres, TLS/ingress, multi-cluster), which may increase time-to-production.
- Operator responsibility for secrets: DB credentials, API keys, and TLS must be managed externally (Kubernetes Secrets, Vault, etc.), adding configuration work.
- Community chart variance: Third-party Helm charts exist and may be outdated—use the official chart to reduce surprise issues.
- Occasional UI/API issues: Some users report intermittent dashboard or API bugs after upgrades; plan for monitoring and quick patching or rollbacks.
Final Thoughts
Flipt is a practical choice if you want a self-hosted, Git-centric feature-flag platform that scales and performs for production workloads—particularly when you already run Kubernetes or need strict on-premises control. It covers the common needs of toggles, variants, targeting, and progressive rollouts while exposing monitoring hooks and multiple evaluation APIs.
If you plan to self-host, start small: evaluate using the in-memory mode or a single-node Postgres, deploy with the official Helm chart, and add Prometheus-based monitoring. Invest in secret-management and upgrade/rollback processes, and consider Flipt’s commercial tiers if you need RBAC, audit logs, or vendor support. If you prefer a no-ops SaaS or lack K8s/ops bandwidth, a managed feature-flag provider may be a better fit.