Every release of Kloudfuse has pursued the same goal from a new angle: unify observability, reduce its cost, and give engineering teams faster answers. Kloudfuse 3.0 is the most complete expression of that goal so far.
OverviewA platform, release by release
It helps to see 3.0 in context — as the third step of a deliberate progression, each release adding a layer while keeping the same architectural core.
1.0The data lake foundation
In our inaugural release, we focused on a robust platform anchored by an Observability Data Lake. Recognizing the fragmentation in the market and the toil developers face with manual issue resolution, we launched with metrics and logs in a single platform — with the vision of integrating additional streams as we expanded.
Early client feedback confirmed that correlating metrics with underlying logs significantly reduced mean time to detect. We also built the data lake for private deployments managed by a control plane, addressing a real gap: organizations frustrated by the high cost of SaaS solutions, vendor lock-in, and security and compliance concerns.
2.0APM & OpenTelemetry
In our second release, we launched Application Performance Monitoring and fully integrated OpenTelemetry to create a robust troubleshooting platform. Customers could now correlate trace data with logs and metrics — a unified solution for both developers and SREs, enabling faster detection and resolution as microservice architectures made distributed tracing essential.
We also integrated AI-driven intelligence, leveraging algorithms such as DBSCAN and SARIMA for smarter insights, and introduced Advanced Services Monitoring (ASM) with eBPF for deeper, holistic visibility into application performance.
3.0Full-stack MELT
Kloudfuse 3.0 builds on those successes. Using our Metrics, Events, Logs, and Traces (MELT) framework, we added frontend observability — Real User Monitoring and session replays for web and mobile — and Continuous Profiling for code-level analysis, helping teams identify low-quality code and protect reliability.
With far more data now flowing through the platform, two initiatives followed: deepening intelligence to extract insight from full-stack data, and controlling the cost of that data.
- Cardinality analytics and roll-ups for metrics and traces — real-time insight into data volumes while reducing storage and processing cost.
- Archival and hydration — retain logs for compliance at a lower cost, rehydrating when needed.
- ARM / Graviton support — the most cost-effective platform in many cloud regions.
- Enterprise controls — enhancements to RBAC, SSO, security certifications, an enterprise service catalog, and selective data isolation based on customer attributes.
On the AI front, 3.0 added Prophet for anomaly detection and forecasting — managing irregular time series with missing values, such as gaps from outages or low activity, for less tuning and better forecasts even with limited training data — and K-Lens, which uses outlier detection across thousands of attributes to accelerate debugging with clear visualizations.
What's nextLooking ahead
We're energized by the feedback from customers already using these capabilities, and we look forward to learning from more developers as they explore the platform. The throughline remains the same: one unified data lake, deployed in your cloud, doing more with every release.