Aperture by Sparvia · Observability & Monitoring
Every call, explained.
When a call degrades, the carrier, Microsoft and your network each prove it wasn't them — and the ticket dies as "no fault found." Aperture watches the SBC leg and the Teams leg of the same call and tells you why.
Aperture watches the same call end to end — and tells you why.
It correlates the SBC leg and the Teams leg of the same session, lines them up against carrier, trunk and client-side quality, and returns a plain-English verdict with the evidence attached. The blame game ends because there's finally one account of the call that everyone can see.
A call degrades. Three teams each prove it wasn't them.
The audio breaks up on an important call. The carrier shows clean trunks. Microsoft shows a healthy tenant. Your network team shows green dashboards. Everyone is right about their own slice — and the ticket quietly dies as "no fault found," because no one was looking at the whole call at once.
Drag to reveal — or use the arrow keys.
Experience Aperture in Motion.
Seven capabilities, ordered the way we sell them — the verdict that ends the blame game first, the governance that lets you run it at scale last. Pick one to see what it does.
Stop guessing whose fault it was — Aperture tells you.
Open any call and get a plain-English verdict, a confidence score, and a suggested fix — not a wall of metrics to interpret yourself. Every verdict links the exact CDRs and quality metrics it reasoned from, so it reads as evidence, not a black box. A thumbs-up / thumbs-down lets your engineers tune it over time.
Watch the fault travel through your network, live.
Carriers, trunks, SBCs and Teams / Genesys rendered as one health-coloured map. When something breaks, the failing path lights up so you can see where the trouble is before you read a single log. Click any node for its own mini-verdict and a jump straight into the incident.
From alert to full story in one click.
Click the incident banner and the whole timeline opens: when it started, which anomaly triggered it, who was paged, and when it escalated. The exact metrics that proved the fault sit right there — no pivoting between four tools to reconstruct what happened.
Every call, searchable.
A live, filterable list of the day's calls — Failed, Degraded, Good — with a plain-English banner summarising the dominant failure pattern so you're not the one spotting it. Export to CSV when you need to take it to a carrier or a change board.
Type a name, get their call history.
Your help desk pastes a person's name and immediately sees their session history, quality trend and Wi-Fi usage. The 'is it just me?' ticket becomes a ten-second answer instead of an escalation.
The right people get paged automatically, in the right order — with proof.
A P1 pages your CIO, senior manager, our principal consultant, the ops lead and the help desk immediately. P2 starts smaller and auto-escalates after 30 minutes if unacknowledged; P3 and P4 taper further. It's the vendor-accountability story made concrete — and auditable.
Point your SBC at us, get an ingest endpoint in seconds — and manage every customer from one pane.
Onboarding, roles and governance live in one place: your syslog ingest host with a ready-to-paste SBC config, named roles scoped to the systems each person runs, a full audit trail, and an MSP switcher that re-scopes the whole app into any sub-tenant. SSO with enforced MFA throughout.
What you're seeing: The verdict card that resolves in when you open a failed call: cause, confidence, and the evidence chips behind it.
Enterprise-grade by default.
The controls your security team asks for on the first call — not a roadmap slide. Earned certifications, and the specifics behind them.
Entra ID SSO + enforced MFA
Every sign-in goes through your identity provider with multi-factor challenged, not optional. No local passwords to leak.
Tenant isolation
Row-level security separates every tenant's data at the database, so one customer's calls can never surface in another's view.
Private ingestion
Bring your telemetry in over Private Link / private networking — your call data never has to traverse the public internet.
PII masking by role
Numbers, names and recordings are masked based on the viewer's role. A help-desk viewer sees less than a principal engineer.
Full audit trail
Logins, MFA challenges and every configuration change are written to an immutable, SOC 2-flavoured audit log.
Regional data residency
Choose Malaysia or Singapore residency — or run the whole platform on-premises for regulated, sovereignty-first buyers.
Choose your altitude.
Four tiers, from SBC-only monitoring to a fully sovereign on-premises deployment. Every tier is priced per monitored session.
| Capability | Basic | Advanced | Enterprise | On-Premise |
|---|---|---|---|---|
| Ingestion | ||||
| SBC syslog + CDR (AudioCodes / Ribbon) | ||||
| Microsoft Teams / M365 connector | ||||
| Zoom & Genesys Cloud connectors | ||||
| Private Link / private ingestion | ||||
| Correlation & AI | ||||
| Deterministic root-cause | ||||
| AI Root-Cause Verdict + confidence | ||||
| SBC-leg ↔ Teams-leg correlation | ||||
| Verdict tuning (👍 / 👎 feedback) | ||||
| AI Inference | Public | Public | Azure AI Foundry | Customer Hosted |
| Dashboards | ||||
| Live health overview | ||||
| Voice topology map | ||||
| Session Explorer + CSV export | ||||
| Help-desk user lookup | ||||
| Data retention | 7 days | 30 days | 12 months | Custom |
| Alerting & Escalation | ||||
| Threshold alerts (Teams / email) | ||||
| ITSM / webhook routing | ||||
| Escalation Matrix (P1–P4) | ||||
| Security & Governance | ||||
| Tenant isolation (row-level security) | ||||
| Entra ID SSO + enforced MFA | ||||
| RBAC named roles | Basic | Basic | Full + custom | Full + custom |
| PII masking by role | ||||
| Audit trail | ||||
| MSP multi-tenancy (sub-tenants) | ||||
| Geo-redundant | Azure Regions | Customer-managed | ||
| MACC eligible | Eligible | |||
| Data residency | Shared region | Shared region | MY, SG, AU or Tokyo | Your data centre |
| Support | ||||
| Support | Community | Business hours | Priority 24×7 | Priority 24×7 + named engineer |
| Onboarding | Self-serve | Guided | White-glove | Custom deployment |
Priced per monitored session — you pay for activity, not headcount.
The things every evaluator asks.
Point your SBC's syslog and CDR feed at the ingest endpoint we give you, consent one Microsoft app for the Teams side, and you'll see correlated insights within the hour. No agents to roll out, no appliance to rack.
Yes — AudioCodes and Ribbon (SBC 1000, SWe Edge, Edge 8100 and similar) via standard syslog and CDR. If your SBC can emit syslog, it can feed Aperture; we hand you a ready-to-paste config snippet during onboarding.
Aperture correlates the SBC leg and the Teams call record of the same call. That's the whole point: instead of a carrier view and a Microsoft view that never meet, you get one account of the session that spans both.
Every verdict links the exact CDRs and metrics it reasoned from, carries a confidence score, and has 👍 / 👎 tuning. It's evidence-first — you can always open the proof behind the answer — so engineers can trust it and correct it, rather than take it on faith.
Tenant isolation with row-level security, Entra ID SSO with enforced MFA, PII masking by role, and a full audit trail. Enterprise data residency covers Malaysia, Singapore, Australia or Tokyo; or run the platform entirely on-premises in your own data centre if your regulator requires it.
Pricing is per monitored session — you pay for activity, not per seat or per engineer. Tiers run Basic through On-Premise; exact figures depend on your volume, so we quote after a short scoping call.
Yes. Aperture is multi-tenant with per-tenant RBAC and an organisation switcher — drop into any sub-tenant and the whole app re-scopes to that customer, then step back out to the parent. One pane for every account you run.
Yes — the On-Premise tier runs the platform in your data centre with customer-hosted inference for the AI verdict, for regulated and FSI buyers who can't send call data anywhere. Same features as Enterprise, fully sovereign.
See your next outage before your users do.
Point one SBC at Aperture and consent one Microsoft app — you'll have correlated, explained call quality within the hour.