Insights
A backend analytics engine you deploy into your own AWS — it ingests your domain events and computes six proven analysis kinds (funnels, cohorts, correlations, and more) with background workers.
Six analytics patterns. Zero SQL.
Funnels, cohorts, correlations and more — every analytics pattern your team needs, point-and-click on the same event stream. Add a new event type once; every analysis can use it instantly.
Standard Analysis
Deconstruct single events across multiple metrics to understand immediate impact.
Funnel Analysis
Map the conversion journey through sequential steps to identify friction points.
Correlation
Discover relationships between distinct events and measure their statistical significance.
Cohort Analysis
Track behaviour patterns of specific user segments over prolonged timeframes.
Comparative
Compare metrics across time periods to identify trends and measure period-over-period changes.
Event Sequence
Analyze ordered patterns with complex temporal constraints and loops.
Don't just watch the data. Move on it.
Each pattern above can fire a trigger — a threshold alert, an anomaly flag, or a scheduled digest — routed through NotificationHub so the right team hears about it on the channel they actually read.
Threshold alerts
Define a line — payment failures over 50 per hour, GMV down 20% week-on-week — and the right team hears about it the moment it crosses.
Anomaly flags
Sequence patterns like rapid cancellations, repeated payment failures, or listing-rejection spikes flag the affected user, brand, or order automatically. No dashboard required.
Scheduled digests
Send any insight on a cadence — daily ops summary, weekly executive snapshot, monthly board roll-up — through the channel each team actually reads.
Three paths. One engine.
Engineers, product teams, and your customers each get the consumption mode that fits them — pointed at the same event stream, the same snapshots, the same source of truth.
Define in code
Analytics-as-code via the private NuGet library. Insight definitions live in your codebase, version-controlled in Git, deployed via CI/CD.
// Pin to Git, deploy via CI/CD
var r = new CreateFunnelInsightRequest
{
Name = "Activation funnel",
Steps = [...],
Metrics = [...],
};
await insights.CreateAsync(r);Build in the dashboard
A React admin UI. Pick a pattern, point at events, name your insight. No SQL, no redeploy, no engineer in the loop.
Embed as widgets
Drop-in React components pre-wired to your insights. Number tiles, sparklines, funnels, retention grids — render any insight in your customer-facing app with one tag.
Browse widgetsFrom raw events to revenue insight.
Six analysis kinds
Standard, Funnel, Correlation, Cohort, Comparative and Event Sequence, each with its own manager to create, update and read results.
InsightKind.EventSequence Computed in the background
SQS workers compute snapshots on each insight’s own EventBridge schedule, reading the events EventHub already logs, in your account.
AddKanjectInsightWorkers Backfill history
Compute an insight over events from before it existed, in daily, monthly, quarterly and yearly phases. Failed tasks can be retried.
CreateBackfillAsync() Threshold alerts
Set an operator and a line, per group if you like. Crossing it publishes an EventHub event, and so does coming back, so any service can act.
insight_threshold_breached Anomaly detection
Flag a value that strays from its last 30 days, at Low, Medium or High sensitivity. It publishes an event when it starts and when it clears.
AnomalySensitivity.High Drill into a number
Fetch the entity keys behind any aggregate, so a spike leads straight to the records that caused it.
GetInsightDrilldownDataAsync Inside your service.
On Lambda or in any container.
Insights is a .NET library, not a service you call. Add the package, register it in Program.cs, and it runs inside your own service, in process. Ship that service as a Lambda function or as your own container on ECS, EKS or App Runner — the module goes wherever your code goes.
- ✓ In process: your code calls the module directly — no sidecar, no extra network hop
- ✓ Built from your code: your image or Lambda package, patched on your schedule
- ✓ Configured in
Program.cs, with values from your own settings
Kanject.Program.cs, and exposes a typed API the rest of your service can use immediately.
From the team — about this module.
Building a serverless analytics engine that scales to zero
Most analytics products solve the dashboard latency problem with an always-on OLAP cluster. We solved the same problem differently — Lambda + DynamoDB + S3 Express + EventBridge in your AWS account. Here's the architecture brief.
Six analytics patterns. Zero SQL. Introducing Kanject Insights.
We shipped a click-to-insight engine that supports the six analytics patterns every product team needs — funnels, cohorts, correlations, and more — without writing a line of SQL.
Simplify your cloud
journey today.
Join forward-thinking developers and businesses who trust Kanject to eliminate cloud complexity and accelerate innovation.