How Streamvane delivers Genesys Cloud KPIs in real time
An event-driven reactive processing platform that turns Genesys analytics events into 1,200+ consistent operational KPIs and delivers them to every connector in under a second. Azure-native by heritage, but cloud-portable — run it embedded in your own stack on Azure, AWS or GCP, or as a managed PaaS.
The problem we remove
One operational truth, not a dozen reporting silos
Most estates don't lack data — they have too many copies of it. WFM exports one service level, a BI script computes another, a client report a third, and nobody agrees which is real. Streamvane computes each KPI once, at the source event, and delivers that single definition everywhere.
- Definition — not dashboard — is the single source of truth
- No more meetings spent reconciling whose number is right
- Planner, BI, executive and client all see the same figure
- Measure once, report everywhere — on infrastructure you own
Design principles
Six principles the platform is built on
Event-driven first
Genesys analytics events over an event bus as the primary ingress — not fragile API polling — for sub-second reactive processing.
Configurable reactive processing
1,200+ structural KPI families enabled by configuration, plus client-specific metrics — never a redeploy of the whole stack.
Your cloud, your data
The operational metrics store lives in infrastructure you own and govern, on Azure, AWS or GCP, in the region you choose.
Multi-connector egress
The same computed KPIs reach Power BI, APIs, feeds and message buses — define once, deliver everywhere.
Two operating models
Consume Streamvane as a managed PaaS bridge, or embed it plug-and-play — components and code — inside your own stack.
Production operations
Dedup, watermarking, monitoring, runbooks and housekeeping built in from day one.
Pipeline
The real-time Genesys KPI pipeline, event to delivery
A real-time KPI pipeline in four stages — storage optional — each observable and individually scalable, sub-second end to end on the streaming path.
- 1
Ingress
The ingress module consumes Genesys events from Azure Service Bus with peek-lock semantics.
- Message validation against expected Genesys event shapes
- Deduplication and watermarking for ordering-sensitive KPIs
- Partition coordination for scaled workers
- 2
Reactive Engine
Topic-aware processors emit activities — intermediate KPI building blocks held in buffers until scheduler intervals flush them.
- Weaver correlates related activities within time windows
- API schedulers poll Genesys REST where event coverage is incomplete
- Separate scheduler paths for the metrics store, Power BI and feed export
- 3
Storage — optional
Optional, and enabled only when you need it: consolidated metrics persist to a metrics store by metric family for downstream reads, history and reconciliation. Pure streaming consumers can deliver KPIs without persisting them at all.
- 4
Delivery
Connectors attach to computed and stored KPIs.
- Power BI: batch push to streaming dataset endpoints
- REST: authenticated HTTP query APIs over the metrics store
- HTTP feeds: timer-driven contractual payload exports
- RabbitMQ: FIL bridge processors publishing to exchanges
Two ways to run it
Managed by us, or embedded in infrastructure you own
Pick the operating model your governance needs. Run Streamvane as a managed PaaS — we operate the Genesys integration and bridge KPIs to your connectors — or embed the architecture, plug-and-play down to components and code, inside your own stack. When embedded, ingest, reactive processing and storage all run in your own subscription and region, on Azure, AWS or GCP; connectors are the only components that reach outward, and secrets never leave your vault.
- Managed PaaS: fastest start, agreed isolation and residency controls
- Embedded: full ownership on Azure (native), AWS or GCP
- Secrets in your Key Vault — no credentials in config files
- Data residency in the cloud region you select
Architecture lens
Why event-driven reactive processing beats the alternatives
Every other approach forces a trade-off: speed, breadth, contact-center accuracy or ownership — pick a few and lose the rest. An event-driven reactive core is the only design that refuses the compromise. It reacts the instant Genesys emits an event, computes each KPI once with real contact-center semantics, and fans the same number out to every destination — inside infrastructure you own. That is the difference between a feed and a platform.
| Approach | Where it breaks down | Streamvane reactive core |
|---|---|---|
| API polling scripts | Fragile and rate-limited, with seconds-to-minutes lag and gaps on every Genesys API change | Event-driven ingress reacting in under a second, resilient to schema drift |
| Hosted BI connector only | A single vendor-hosted sink — your KPIs live in someone else's tool, not your estate | Multi-sink delivery from one core, running in your own cloud |
| Generic iPaaS / ETL | Moves data but has no Genesys semantics — every KPI is re-derived by hand and drifts out of sync | Topic → activity → KPI mapping, defined once and reused everywhere |
| Nightly batch / data dumps | Yesterday's numbers — useless for live floor decisions and real-time alerting | Continuous computation; supervisors and systems act on the same instant |
| SI bespoke build | Slow, expensive and non-reusable; each destination is a fresh project that no one else can maintain | Accelerator IP plus configuration — new KPIs and connectors are config, not rebuilds |
Same KPI, computed once, delivered everywhere in under a second — the compromise other approaches force simply isn't there.
1,200+ structural KPIs, plus anything bespoke — as configuration
A standard ~1,000-agent deployment ships with 1,200+ structural KPIs derived straight from Genesys topics — and any metric unique to your business is modelled on the same engine as configuration, never a fresh ingestion build per consumer. The platform itself is cloud-portable and language-friendly: consume it as a managed PaaS where we operate the integration, or embed the whole architecture as a plug-in inside your own systems — on Azure, AWS or GCP, in the languages your teams already maintain.
- Runs on Azure, AWS or GCP — your cloud, your region
- Integrates with .NET, Java, Ruby, Go and Rust codebases
- Any custom KPI modelled to your definition, delivered in real time
- Consume as managed PaaS, or embed the architecture in your own stack
- Structural KPI families enabled by configuration
- Connector and scheduler rules tuned per destination
Walk through the architecture for your estate
Bring your cloud and Genesys context to a technical discovery, and we'll map the reference architecture onto your environment — Azure, AWS or GCP. We'll show where ingress, the reactive core and the optional store land in your subscription, which connectors light up first, and whether you run it managed by us or embedded plug-and-play, components and code, inside your own stack.
Streamvane by Tessovia · Azure · AWS · GCP · Your data, your keys