Current release:
v1.0.0-alpha
All features available in the current release.
- Kubernetes Collector - Watches Deployments, ReplicaSets, ConfigMaps, and Secrets with annotation-based filtering
- Execution Fingerprinting - Ignores rollouts where the generation changed but the fingerprint (git SHA) remained identical, filtering out manual
kubectl editnoise - Orphan Detection - Highlights execution events with no matching intent (CI) event
- Deterministic Impact Scoring - Baseline vs impact window comparison with configurable weighted scoring for built-in and custom metrics
- Immutable Analysis Snapshots - Results stored permanently, never retroactively altered
- Confidence Scoring - Secondary score based on traffic volume significance
- Change Ranking - Ranks concurrent changes by likelihood of causing degradation
- Customizable PromQL Templates - Core queries configurable via
config.yamlwith{{ .Services }}and{{ .Duration }}variables - Additional Custom Metrics - User-defined PromQL queries in config, collected alongside core metrics
- Automatic Background Analysis - Worker triggers analysis when impact windows close; polling interval configurable
- Event Retention Policy - Configurable TTL with automatic cleanup
- Intent-Execution Linking - Links CI events to K8s rollouts via a 3-tier confidence ladder:
sha_match(1.0),image_tag_match(0.9, zero-config),image_sha_inferred(0.85). All links include explainablereasonmetadata. - Timestamp Guardrail - Prevents
ChangeEvents with future timestamps from entering the analysis workflow; marks them invalid for auditability - Contextual Deep Linking - Dynamic clickable links to external systems (Git, CI/CD) generated from event metadata via configurable
LinkTemplates - Full UI - Dashboard timeline, service analytics, metric preferences, search, namespace/type filters
Theme: Complete the Foundation
| # | Item | Status |
|---|---|---|
| 1 | Analysis status badges in event timeline | ✅ Done |
| 2 | Live image diffing on deployment events | ✅ Done |
| 3 | Deployment Story timeline - intent + execution as a single unified card (CI Build → Image Push → K8s Rollout → Impact Score) |
✅ Done |
| 4 | Multi-metric UI toggle in service analytics | ✅ Done |
| 5 | Helm chart - helm install valiant valiant/valiant, listed on Artifact Hub on release |
✅ Done |
| 6 | GitHub Actions example configs - drop-in workflow snippets for CI/CD event injection | ✅ Done |
Theme: General Availability Status: Planned
Production-ready. OpenShift-capable. The version teams can confidently run in a real cluster.
Cluster compatibility
- OpenShift SecurityContextConstraints - full support for restricted SCC environments
- OpenShift RBAC manifests - pre-built Role and RoleBinding manifests alongside standard Kubernetes ones
UI & operations
- Service Health Pulse - traffic-light indicator per service derived from last known impact score ✅
- Concurrent Group Visualization - highlight services that deployed within the same time window
- Event search by Git SHA and arbitrary metadata values
- Markdown/URL rendering in event metadata fields ✅
- Retention settings exposed in the UI (no more editing
config.yamlfor TTL)
Onboarding
- Demo seed environment - realistic multi-service deployment story out of the box; populated dashboard in under five minutes
Theme: Deeper Insights Status: Planned
Make Valiant stickier for teams using it daily.
- Manual verification tags on snapshots - annotate results with
confirmed root cause,false positive, orrolled back - Blast radius visualization - UI surface for ConfigMap/Secret → affected workloads (data already exists in the API)
- Per-service configurable impact thresholds - LOW/MEDIUM/HIGH tunable per service
- Refined likelihood scoring for ConfigMap and Secret changes vs. image rollouts
- Filtered event export as JSON - for incident reports and external tooling
Theme: Ecosystem Integrations Status: Planned
Connect Valiant to tools teams already use. This is the discovery milestone - Grafana plugin registry, Slack, ArgoCD, and GitHub put Valiant in front of new users without any marketing effort.
Observability
- Grafana annotation injection - automatic deploy markers on all existing dashboards
- Grafana panel plugin - listed in the official plugin registry
Notifications
- Slack and generic webhook notifications on MEDIUM/HIGH impact results
Event sources
- Git Collector - direct collection of tags, releases, and merges from Git repositories
- GitHub webhook receiver - push, tag, and release events with structured metadata
- GitLab webhook receiver
- ArgoCD notifications integration - covers GitOps clusters with no CI pipeline
Theme: Proactive Awareness Status: Vision
Shift from "what happened after my deploy" to "something looks wrong right now."
- Live rollout health scoring - real-time metric trend updated as the impact window fills, not only at the end
- Rollback recommendation - when HIGH impact is detected, surface affected metrics and a direct link to the rollback action in ArgoCD (human decision only, full context surfaced immediately)
- Alertmanager enrichment - Valiant enriches Alertmanager alerts with recent changes to the affected service
- Pre-deployment risk estimation - based on historical snapshots per service, estimate risk profile before a new version ships
Keeping this honest. Valiant is a focused, single-cluster change impact tool.
- Multi-cluster and multi-environment management
- Authentication / SSO for the Valiant UI itself
- Distributed event bus (Kafka, NATS)
- Service mesh dependency and proximity scoring
- Full postmortem workflow and incident management