Event-driven Triggers
The events that matter rarely live in one stream. A position breach often occurs as the result of a related asset's price moving suddenly in the market. A counterparty downgrade is often triggered when an exposure crosses a threshold. 3forge detects composite conditions across multiple live streams and resting state, and fires automated actions, notifications, or downstream pipeline stages the moment the pattern matches.
The Challenge
Useful Triggers Are Composite, and Most Systems Cannot Express Them
Single-stream alerting is easy and noisy. The valuable signal almost always comes from a combination, an event in one feed plus a state in another plus a time window. Most orchestration tools either flatten everything into one ordering and lose context, or force each rule into its own bespoke pipeline.
Cross-stream correlation is hard
Joining a high-frequency price stream against a slower position update stream within a defined time window, and only when both conditions hold, is beyond what most pub-sub systems can express without custom code.
Alert fatigue and noise
Single-condition alerts produce thousands of irrelevant notifications per day. Operators stop reading them, and the actual signals get lost in the noise.
No state across events
Many useful patterns require remembering recent state, the last five quotes, whether an order has been working for more than thirty seconds, the running average over a window, and stateless event processors cannot express this cleanly.
Brittle action wiring
Once a trigger fires, the action, a kill switch, a Slack post, a Kafka event, a database update, is wired by hand for each rule. Adding a new action means redeploying the rule.
The 3forge Approach
Pattern Detection Native to a Real-Time Runtime
3forge models every stream as a real-time table and every trigger as an AMIScript expression over those tables. Composite conditions, windows, joins, and stateful aggregations are first-class language constructs. When the condition matches, the runtime invokes any handler, from publishing an event onto Kafka to calling a downstream REST endpoint, all configured declaratively. Rules are versioned, observable, and replayable from the column store.
Key Capabilities
What You Can Build with 3forge
Composite Condition Detection
Express triggers that require multiple streams to satisfy related conditions inside a defined time window, in a single AMIScript expression.
Windowed and Stateful Rules
Use sliding and tumbling windows, persistent counters, and running aggregates to express patterns that depend on recent history.
Pluggable Action Targets
Route trigger output to alerts, Kafka topics, REST endpoints, database writes, or downstream 3forge pipelines through configuration.
Rule Composition
Build complex triggers from simpler named rules, reuse them across applications, and version them like code.
Replay and What-If
Replay any historical period through current rules to backtest, tune thresholds, and demonstrate behaviour to auditors before activation.
Live Rule Observability
Inspect every rule's match rate, false-positive ratio, and downstream action latency in the workbench, with drill-down to the firing event chain.
Get started
Ready to see it live?
Book a 30-minute demo with a 3forge solutions engineer and watch a composite trigger fire across live streams in real time.