Patent-pending fingerprinting organizes your logs into patterns at ingest, compressing storage up to 20× and surfacing anomalies before you search. FuseQL goes beyond search with built-in anomaly detection, forecasting, and analytics operators.
Kloudfuse Logs · a multi-stage FuseQL query over fingerprinted patterns
Patent-pending fingerprinting splits every log line into static patterns and dynamic values before storage. Investigation starts from patterns, not raw text.
Anomaly detection, outlier detection, and forecasting built into the query language. Search across labels and facets in one expression.
Logs join to spans, metrics and RUM in a single query. No cross-product context switching, no separate billing SKU.
Ship logs from whatever you run today. Kloudfuse auto-detects facets and generates fingerprints at ingest. Log processing pipelines let you transform, enrich, filter, and route logs before storage — with stream-level rate control to cap noisy sources.
Every log line is split into a static pattern and dynamic values. The pattern is stored once with occurrence counts. Dynamic values are stored separately. Storage cost stops scaling with log verbosity.
Log patterns — raw lines grouped into fingerprints, with volume per pattern
FuseQL goes beyond search. Anomaly detection, outlier detection, and forecasting are native operators, not add-ons. Search across labels and facets in a single expression with sub-second response.
# unique errors per minute, with a one-hour seasonal forecast * | timeslice 60s | count_unique(@error) by (_timeslice) | predict(_count_unique) by 60s, model=seasonal, seasonality=hourly, forecast=3600s
Group by fingerprint to see what's noisy. Drill into a specific log to see its trace context, span ID, and extracted facets. Switch views without switching tools.
Log search — filtered live tail with the matching facets in the sidebar
Logs live in your VPC, under your keys. Set custom retention policies per stream, archive to S3 for long-term compliance, and hydrate on demand. Access controls and scrubbing rules are configured by your team, not by a vendor pricing model.
| Tier | Retention | Storage | Typical use |
|---|---|---|---|
| Hot | 7–30 days | NVMe / SSD | Frequently accessed — alerts, dashboards, incident debugging, security investigations |
| Warmcoming soon | 1 month – 1 year | S3, pulled at query time | Less frequently accessed — incident debugging, security investigations |
| Cold | 1–7 years, you choose | Object store (e.g. S3 Intelligent-Tiering), hydrate on demand | Compliance retention; hydrated as needed for security and compliance reviews |
FuseQL is Kloudfuse's query language for log analytics that open standards don't cover — parsing, aggregation, pattern extraction and scheduled search. It runs against the same lake as PromQL, LogQL and TraceQL, under the same access controls.
At ingest, every log line is split into a static fingerprint and dynamic values. The static fingerprint is stored once with occurrence counts. Dynamic values are stored separately. Storage cost stops scaling with log verbosity.
FuseQL. Anomaly detection, outlier detection, forecasting, and arithmetic operators are built into the language. FuseQL is the primary log query language in Kloudfuse, with LogQL still supported for existing workloads.
Inside your VPC, under your keys. Kloudfuse runs log ingestion, parsing, and storage in your customer-owned data plane. PII scrubbing happens at ingest, before storage.
Predictable pricing on one axis: terabytes ingested, across every signal. No per-host charge, no per-seat tier and no cardinality premium. Fingerprinting means verbose logs do not drag the cost curve up with them.
Yes. All telemetry lives on the same data plane. Click a trace ID in a log line to jump to the corresponding trace, or pivot from a metric anomaly to the underlying log patterns. Correlation happens through shared identifiers like trace ID, span ID, and facets — no cross-product context switching, no separate billing SKU.
Yes. Kloudfuse supports archival to S3 with configurable rules per source and label. When you need archived logs for an investigation or audit, hydrate them on demand — they come back fully indexed and searchable, with fingerprints and facets intact. In 4.0+, hydration runs in parallel with pause/resume support.
See how Kloudfuse splits log lines into static patterns and dynamic values, compresses storage, and turns fingerprints into first-class investigation surfaces.
Read the guideSmarter log comparisons, structured JSON parsing, and faster matching — all built into the query language, no pipelines required.
Read the blogRetention model, query language, ingestion cost, and data residency. A side-by-side breakdown for teams evaluating both.
See the comparisonThirty minutes on your telemetry. The cause, before the call ends.