Skip to content

// voip monitoring for msps

Twenty customers, twenty phone systems, one place to watch them.

Managing voice for other people is a different problem from managing your own. You need monitors grouped by customer, alerts routed to whoever holds that contract, a status page each customer can check without emailing you, and a record of exactly what the carrier was doing at 02:14 when they insist nothing was wrong.

Start monitoring freeread the docs

tenancy

Grouped by customer, from check to status page

A customer group is not a tag. Monitors, alert rules, status pages and SIP credentials all hang from it, and a credential scoped to one customer cannot be attached to another customer's check.

Access is org-wide rather than per customer. Owners and admins manage billing, probes, alert routing and credentials; members see every customer and manage monitors.

  • Alert rules per customer: the engineer on that contract, not the whole team
  • One public status page per customer, 90 days of uptime, your title and logo on MSP
  • SIP credentials shared per customer: REGISTER and call-quality checks on one PBX use one record, so a password rotation is one edit
  • A private probe with its own region label runs only the checks you point at it, so it can sit inside one customer's network
  • Per-monitor pricing, so a one-trunk customer costs one monitor

customers

acme-telecom12 monitorsup
noordzee-logistiek6 monitorsup
van-dijk-advocaten4 monitorscert 9 d
helder-zorg9 monitorsup

31 monitors · 1 warning · 0 down

evidence

Something to put in the carrier ticket

The hardest part of managing someone else's voice platform is proving where the fault sits. Two probes settle it: one outside on the public internet, one inside next to the SBC.

When the external probe records 503 from the carrier edge at 02:14 while the internal probe shows the PBX answering OPTIONS in 3 ms throughout, the conversation with the carrier starts from transactions and timestamps rather than from a customer's memory of a bad morning. Every check result is retained with its response code, round-trip time and the probe that recorded it, and SIP checks keep the messages themselves, so the ticket can quote the 503 with its headers instead of describing it.

why a generic uptime tool can't do this

questions

Straight answers

What should an MSP monitor on a customer VoIP platform?

At minimum: SIP reachability of the SBC or PBX edge with OPTIONS, the registration path with a dedicated test extension, TLS certificate expiry on the SIP TLS port, and the NAPTR and SRV records that route calls to the platform. Add an RTP test call where call quality complaints are frequent or where the contract carries a quality commitment. That set covers signalling, authentication, encryption, routing and audio, which is every layer between the customer and a completed call.

Can each customer get their own status page?

Yes, on the Pro and MSP plans. Monitors are grouped per customer, and each group can publish a public status page at app.sipsentry.net/status/your-slug, with 90 days of uptime history for the monitors you choose to show. On the MSP plan the page carries your own title and logo and drops the Powered by SIP Sentry footer, so what customers see is their voice platform rather than your tooling. Customers who can self-serve the answer to "is it us or is it them" generate noticeably fewer tickets.

How do alerts get to the right engineer?

Alert rules are scoped org-wide, per monitor, or on the MSP plan per customer, and each rule picks its events: down, recovered, degraded or certificate expiring. Point a customer rule at the channel of the engineer who holds that contract and they get that customer, not the whole book. Rules add up, so the same alert can reach several channels: typically Slack for the team and SMS for whoever holds the out-of-hours phone. Channels are email, HMAC-signed webhooks and Slack, plus Telegram and SMS through your own bot and SMS gateway.

Does monitoring from outside prove the problem is the carrier?

It proves a great deal of it. When an external probe records 503 responses from the carrier edge while a probe inside the customer network shows the PBX answering normally, the direction of the fault is no longer a matter of opinion. Every result is timestamped and retained, so a carrier ticket can cite specific transactions with response codes and round-trip times instead of a customer recollection of roughly when calls got bad.

What does it cost per customer?

Pricing is per monitor, not per customer or per seat, so a customer with a single trunk costs one monitor and a multi-site customer costs as many as you choose to watch. The Pro plan covers 50 monitors at €29 per month with customer groups and status pages; the MSP plan covers 250 at €99 per month, routes alerts per customer and white-labels the status pages with your own title and logo. Three monitors are free forever, which is enough to trial the approach on a single customer before committing.

Point one check at it and find out

Three monitors free, forever. No card, no sales call, first result inside a minute.

Start monitoring freesee pricing