Real-time dashboards
Sub-second queries at scale
Stream events via the collector API and read them back in under 300 ms p95 — full fidelity, no pre-aggregation, no stale tiles.
Real-time event analytics
From raw events to answered questions in under a second.
Nimbus streams product, pipeline and infrastructure events into one warehouse-native layer — sub-second p95 queries, zero sampling, SQL on day one.
SOC 2 Type II · GDPR-ready · 99.99% ingest uptime
Events · last 12 hlive
Events / sec48,210+12.4%
Funnels / day1,284+3.1%
Error rate0.42%+0.08 pp
Hourly event volume rose from 41,000 to 63,000 events per hour.
p95 query 240 ms · 12 h window · UTC
Platform
Analytics your engineers won't fight
One event pipeline feeding dashboards, funnels and your own SQL — with no data leaving your stack.
Funnels & segmentation
Funnels that match your schema
Compose multi-step funnels on any property with windowed conversion, cohort breakdowns and segment diffs — no rigid naming conventions.
Warehouse-native SQL
Your warehouse, your SQL
Models sync to Snowflake, BigQuery and Postgres as dbt-compatible tables, so analysts query the same rows in raw SQL.
How it works
From snippet to insight in three steps
01
Instrument
Drop in the 11 kB SDK or a server library. Autocapture covers pageviews, clicks and identity in one snippet.
02
Stream & validate
Events hit the edge collector, get validated against your tracking plan and deduplicated before a single row is written.
03
Query
Dashboards, funnels and SQL run against columnar storage with p95 latency under 300 ms at billions of rows.
Customers
Trusted by data teams who count events
"We replaced three bespoke dashboards and a nightly Airflow rollup. Funnel answers that took an engineer a day now take a PM two clicks."
Head of Data Infra, Larkfield
"The ingest pipeline shrugged off Black Friday: 2.1 billion events in 24 hours with zero backpressure and no dropped payloads."
Staff Platform Engineer, Northwind Logistics
"Warehouse-native sealed it. Our dbt models became queryable product analytics without moving a byte."
Analytics Lead, Corvid Systems
Pricing
Metered by events, not by seat
The first 1M events every month are free — on every tier, forever.
Save 20%
Starter
$0 /mo
- Up to 1M events / mo
- 2 seats · 30-day retention
- Core dashboards
- Community support
Pro
Recommended$49 /mo
billed monthly
- Up to 25M events / mo
- 10 seats · 13-month retention
- Funnels, cohorts & SQL editor
- SSO + priority support
Enterprise
Custom
- Unlimited events & seats
- Dedicated ingest cluster
- SOC 2, DPA · 99.99% SLA
- Audit logs & VPC peering
FAQ
Technical questions, direct answers
How does Nimbus ingest events without sampling?
The edge collector buffers micro-batches, validates each payload against your tracking plan and writes full-fidelity rows to columnar storage. Dashboards always query 100% of events.
Which warehouses does Nimbus sync to?
Snowflake, BigQuery, Redshift and Postgres. Models are materialized as dbt-compatible tables and refreshed on the schedule you define, from every 15 minutes to nightly.
What does the SDK do to page performance?
The browser snippet is 11 kB gzipped, sends asynchronously via sendBeacon and never blocks the main thread. Our p75 impact on LCP across monitored sites is under 20 ms.
Is data retention configurable?
Yes — from 30 days on Starter to custom hot/cold tiers on Enterprise, with automatic Parquet export to your own bucket when data ages out of the hot layer.
How is billing metered?
By ingested events, computed monthly. Events that fail tracking-plan validation are never billed, and the first 1M events each month are free on every tier.
Ship analytics your engineers trust
1M free events a month. No credit card, no sales call — your first dashboard is ten minutes away.