Skip to content
NETWORK OKASR 67.3%PDD 1.84sCPS 2,412
VICICarrier
SLA · SERVICE LEVEL AGREEMENT

99.99% uptime. Real credits when we miss it.

Trailing-90-day uptime has been 99.998% since we started publishing it. Our contractual floor is 99.99% and our credit schedule actually pays. No 'credit capped at 5% of monthly spend' fine print, no force-majeure catch-alls, no support-hour exclusions.

S.01 / COMMITMENTSNETWORK + VOICE QUALITY

The four numbers we write down.

An SLA that only promises uptime is useless — a 99.99%-up trunk with 3% ASR and 4-second PDD is a broken trunk. We commit to all four axes that actually matter to a dialer floor.

Uptime — 99.99%

Measured per-POP, trailing 30 days, excluding customer-side configuration errors only. Scheduled maintenance is announced 72 hours ahead and does not count against uptime.

ASR — US domestic ≥ 67%

Answer Seizure Ratio on US/CA domestic routes, rolling 7-day window per tenant. Falls below floor → credit triggers, investigation starts within 2 hours.

PDD — mean ≤ 2.0s

Post-Dial Delay measured at our SBC, US/CA routes. Tail latency (p95) committed at ≤ 3.5s. Published per-POP on /status.

MOS — ≥ 4.2

Mean Opinion Score on recorded and sampled calls, measured via PESQ. Below 4.0 triggers route-quality investigation regardless of uptime.

S.02 / CREDIT SCHEDULEREAL DOLLARS · NOT 'SERVICE CREDIT'

Miss the floor, pay the bill.

Credits apply to the affected service only (trunk, DID block, or SMS campaign), not your whole account. They're issued as invoice credits by the 15th of the following month — no support ticket required, no 'submit within 5 business days' gotcha. We calculate from our own monitoring and publish the breach in /status.
Monthly uptimeCredit (% of affected monthly fee)Also includes
≥ 99.99%0%
99.9% – 99.99%10%Written incident report
99.5% – 99.9%25%Root cause + remediation plan
99.0% – 99.5%50%Next-month waiver on number fees
< 99.0%100%Termination without penalty
S.03 / SUPPORT RESPONSE24/7 NOC · NAMED ENGINEERS

Tiered response. Humans on both ends.

Severity is declared at ticket open and confirmed within 15 minutes — not re-classified later to dodge the SLA clock. After-hours tickets route to the same NOC engineers who run the network, not a first-line script farm.

SeverityExamplesFirst responseStatus update
SEV-1 · outageTrunk down, all DIDs failing, total ASR collapse≤ 15 min · phone + ticketEvery 30 min until resolved
SEV-2 · degradedPartial route failure, ASR drop > 15%, MOS < 3.5≤ 30 min · ticketEvery 2 hours until resolved
SEV-3 · issueDID provisioning delay, CDR gap, single-call anomaly≤ 2 business hoursDaily until resolved
SEV-4 · questionConfig help, documentation gap, feature request≤ 1 business dayOn meaningful progress
S.04 / EXCLUSIONSWHAT WE WON'T CREDIT

Short list. In plain English.

Customer-side misconfiguration

Wrong trunk IP, missing ACL, expired cert on your SBC. We'll help you fix it — we won't credit for it.

Upstream carrier-of-carrier outages that hit the whole US/global PSTN

Century-level incidents (e.g., a Tier-1 backbone cut). We disclose which events qualified and why on /status.

Scheduled maintenance

Announced ≥ 72 hours ahead via email + portal banner + /status. Always during 02:00–05:00 UTC window. Typically < 30 min actual impact.

DDoS attacks aimed at customer IPs

We protect our SBC perimeter. DDoS aimed at your servers is your incident. We'll work with you on rerouting.

S.05 / HOW WE MEASUREFROM THE OUTSIDE · NOT OURSELVES

Synthetic probes + customer-side data.

Uptime and PDD are measured by ThousandEyes probes in 14 global POPs plus 4 customer-hosted probes we provision for free to any customer spending > $5k/mo. ASR and MOS come from your own CDRs — we don't get to pick which calls count. Monthly reports publish the raw numbers and the methodology. If you want the probes in /your/ data center, ask and we'll ship hardware.
TRANSPARENCY > THEATRE

The uptime number on the homepage is a live measurement, not a marketing claim.

Check it yourself. If you see anything lower than 99.99% trailing-30-day on any POP, tell us and we'll either fix it or credit it — usually both.