siliconindia

The core pillars of cloud-native observability workflows, which include metrics, logs, and traces, are typically collected using disparate tools. Combine this with complex cloud and Kubernetes deployments having numerous events to manage, and troubleshooting and identifying the root cause becomes similar to finding a needle in a haystack.

This is where Kloudfuse comes to the rescue, offering the industry's first unified and scalable observability data lake for all high cardinality data streams. It enables businesses to derive the best value from metrics, events, logs, traces (MELT), and more using real-time alerts, advanced analytics, and automation.

Pankaj Thakkar, co-founder and CTO of Kloudfuse, notes that organizations gain a holistic understanding of what happens within their complex application environments when data sources are combined and analyzed. Previously, when the term 'observability' was not part of IT practitioners' lexicon, companies like Splunk, Datadog, and Google targeted each stream of observability separately. Though each pillar alone provides valuable information, it does not present the complete picture. Recently, with the routine adoption of complex, multilayered, cloud-based infrastructures using microservices and containers, unified observability in enterprise IT has become mainstream.

The Kloudfuse platform has unified data storage purpose-built for high cardinality, high volume, and dynamic observability data. It simplifies the operational side by facilitating customization and optimization of each data type. The platform also correlates data across all streams with respect to time and topology, so only relevant data required for troubleshooting is surfaced. Above all, Kloudfuse enables enterprises to ingest, store, and query all observability streams with minimal resources.

A highlighting aspect of the Kloudfuse platform is that it uses the Apache Pinot columnar storage, which enables clients to use compression techniques like dictionary encoding, double delta, and gorilla encoding combined with selective compression algorithms, lowering storage costs while enhancing query performance. The platform also provides operational ease like SaaS platforms with deployment models and functions like open source, giving users a SaaS-like experience through an optional control plane for managing the data plane's lifecycle.

"We want customers to feel in charge of their data. By removing limitations on observability data, we help them analyze how much data they need to store, process, or consume," says Thakkar.

Using an open standard for ingestion and querying, Kloudfuse can ingest various formats for different streams (Prometheus, FileBeat, OpenTelemetry, DataDog) across many sources (Cloud, common framework, Kubernetes), and support various scenarios of existing and new collectors. Kloudfuse also ropes multiple query interfaces (PromQL, LogQL, GraphQL, SQL). Alternatively, it provides an onboarding suite to facilitate migrating dashboards and alerts from vendor-specific formats (DataDog/Wavefront, etc.) to the Kloudfuse platform.

Kloudfuse's intuitive AI-powered HawkEye and BullsEye analytics engines are proven to take a business's observability data to the next level. These engines generate intelligent auto-alerts, contextualize information across streams, and filter noise for the fastest possible troubleshooting. HawkEye automatically watches for anomalies in critical application signals and detects abnormal behavior. Simultaneously, the BullsEye correlates signals across all observability streams, time, and space for the potential root causes of an anomaly in minutes.

We want customers to feel in charge of their own data. By removing limitations on observability data, we help them analyze how much data they need to store, process, or consume

Through these unrivaled competencies, Kloudfuse has spawned many success stories. In one instance, it aided a logistics business struggling with unpredictable data volume from a Google stack driver. Kloudfuse helped them lower costs and effectively compress logs using fingerprinting technology. It also extracted the facets so they could query the data more efficiently. Ultimately, it enabled their developers to debug issues much faster, reduced the logging bill, and improved the time to resolution.

Kloudfuse continues to be a leading player in the industry by leveraging its platform to unify all types of telemetry streams into one highly scalable, high-efficiency, and high-performance datastore purpose-built for observability and troubleshooting.