Petabyte-scale logging
Orbital indexes distributed logs at ingestion, providing sub-second latency for complex queries without the overhead of maintaining an ELK stack.
Indexing logs for systems built at
Performance benchmarks
We bypassed the JVM and standard indexing overhead to build a custom storage engine for high-cardinality machine data.
Orbital executes full-text searches across trillions of events in under 800ms without relying on warm caches.
Store data in your own S3 buckets. We separate compute from storage so you scale retention independently of your budget.
Stop managing mapping explosions. Our engine scans raw data formats directly, eliminating the 3x storage tax of traditional indexes.
Define granular access policies. Mask sensitive PII or restrict specific log streams based on team identity and environment.
Logs are searchable the millisecond they hit our gateway. No batching delays, no polling, and no 'indexing lag'.
Pipe query results directly into your local workflows. Our API supports structured JSON output for automated alerting and analysis.
The architecture
Orbital bypasses traditional ingestion bottlenecks by decoupling compute from storage, allowing you to query raw logs instantly.
Grant read-only access to your S3 or GCS log archives. Orbital maps your schema automatically without moving your data.
Set granular TTLs and cold-storage tiers. Only pay for the compute you use to query, not for keeping data in a proprietary index.
Execute high-cardinality searches using standard SQL or our optimized CLI. Results return in milliseconds, even on multi-terabyte scans.
Verified deployments
Real feedback from infrastructure engineers managing high-throughput production clusters.
We replaced a 40-node Elasticsearch cluster with Orbital. Our p99 search latency dropped from 30 seconds to under 400ms without changing our ingestion pipeline.
Finally, a logging tool that doesn't choke on high-cardinality metadata. It handles our VPC flow logs without the usual indexing tax.
The CLI is actually intuitive. I can pipe raw JSON into complex aggregations across three regions without leaving the terminal or waiting for a browser refresh.
Orbital is the first tool we've used where the cost stays flat even as our log volume spikes during DDoS mitigation. The decoupling of storage and compute is real.
The regex engine is impressively fast. Queries that used to time out on our old stack return instantly here.
No more managing heap sizes or shards. We just point our collectors at the endpoint and it works.
Query petabytes in sub-second latency. Deploy Orbital as a managed service or within your VPC.