VoIP Voice Quality Monitoring · OnStak Use Case

Fix the call before it drops.

Bad voice quality is the kind of failure a customer hears instantly. When packet loss and jitter climb and 40% of complaints have no cause, every call is a risk. Here is how one grocery chain made its voice network clear again.

40% lower

latency, 200ms to under 120ms

30ms → <0.8ms

jitter

3.2 → 4.1+

MOS score

Store associate on a headset making a clear VoIP call

Story highlights

  • The call path, end to end. Voice monitoring on the SIP servers gave real-time insight across the whole delivery chain.
  • Clarity restored. Latency fell 40% and jitter dropped from 30ms to below 0.8ms, lifting the MOS score from 3.2 to 4.1+.
  • Alerts that reach the team. Webhook-based alerting in team messaging replaced reactive troubleshooting.

The situation

A chain whose calls kept breaking up.

The grocery chain ran voice across its network, but congestion drove packet loss and jitter past the point of clarity, and most voice complaints had no clear cause. With no alerting in place, problems were found by ear. They set out to see the voice path end to end.

The problem

The voice network degraded, silently.

Calls broke up, and the team could not always say why. Three gaps drove it.

Packet loss over threshold

3-5% packet loss during congestion, against a 1% threshold.

Jitter hurt clarity

Jitter exceeding 30ms degraded call quality.

Complaints without cause

40% of voice complaints had no clear root cause, and webhook alerting was absent.

What OnStak did

Watch the whole voice path, in real time.

OnStak monitored voice end to end and pushed alerts to where the team already works.

1

Monitored the SIP path

End-to-end voice path monitoring deployed on the SIP servers.

2

Saw quality in real time

Real-time insight into VoIP performance across the delivery chain.

3

Alerted into team messaging

Webhook-based alerting integrated into the tools the team uses.

4

Turned reactive into proactive

Issues were caught before customers heard them.

What changed

From calls found broken by ear to caught in advance.

Before

The customer heard the fault first.

  • 3-5% packet loss against a 1% threshold.
  • Jitter exceeding 30ms hurt clarity.
  • 40% of complaints had no cause, and no alerting.
After, with OnStak

Now the team catches it first.

  • Latency down 40%, 200ms to under 120ms.
  • Jitter to below 0.8ms.
  • MOS score 3.2 to 4.1+.

The outcome

Anonymized grocery chain · voice

Clear calls, caught before the customer.

40%
Lower latency
200ms to under 120ms.
<0.8ms
Jitter
Down from 30ms.
4.1+
MOS score
Up from 3.2.
Proactive
Alerting
Reactive troubleshooting eliminated.

The idea that holds it together

A clear call is a kept customer.

“Voice quality is heard before it is measured. Observability catches the drop before the caller does.”

Awais Janjua  /  Chief Technology Officer, OnStak

Built on
Splunk observability platform End-to-end voice path monitoring SIP server monitoring VoIP delivery-chain insight Webhook alerting

Coffee? No Slides

No pitch deck. No 47-page proposal. Just a straight talk about what’s broken and what to fix first.

  • Solutions
  • Debrief
  • About Us