Seeing what’s happening across my network
A population of calls rather than a single one, counted and grouped and compared, asked in the words you already used when you thought of the question.
These questions normally get answered by whoever knows the reporting tool well enough to build the query, which in practice means one or two people per company. Everyone else waits for them, or does without.
The gap is rarely the data. It is knowing which field the failure lives in, remembering that the detailed cause is not the same thing as the SIP code, and being fluent enough in the filter syntax to express “grouped by host” before you have lost interest. Butler takes the sentence and does that part.
Pick a question, read the whole exchange
Both of these are unedited machine transcripts, including the parts where the honest answer is a small number.
- Cause 216 indicates the SBC could not allocate the resources needed to handle the call, such as media channels, transcoding resources or call slots.
- The call was terminated with SIP 480 and mapped to NoAnswer.
- One call in 48 hours is very low volume, so this reads as an isolated resource constraint rather than a widespread issue.
The useful part of this answer is the number one. A cause code you have never seen before feels like the start of an outage, and the first thing worth knowing is whether it is a pattern or a single event. Butler also did the translation nobody asked for: 216 is the detailed cause, the call signalled as a SIP 480, and those are two different fields that a search can disagree about.
An illustration rather than a case study: this one was run against synthetic data to show the shape of the answer, because no customer has asked it yet. What it demonstrates is a comparison holding its nerve. Volume fell and answered calls fell with it, and the answer is that nothing changed, which is a harder thing for a report to say than finding a problem.
One call, one number, the right conclusion. “I cannot reach it” and “it is down” feel identical at the moment you notice and lead to completely different afternoons, and the box reporting a minute ago settles which one you are in. Notice what it does not do: having answered, it stops, rather than filling the space with a list of things that might be wrong with your network.
The prompts, and where each one leads
A failure mode, a time window and a grouping, in one English sentence.
You have seen one instance of a cause code and want to know whether it is a pattern or a one-off.
The count over the window, with the calls behind it available to open one by one.
The shape a manager asks in rather than an engineer. One line, two standard metrics, no dashboard to open.
Volume and answer-seizure ratio for the period, in the thread.
Capacity planning, or a complaint that calls are being rejected at busy hour.
Peak concurrency per NAP for the period, so you can see which trunk is near its ceiling.
Something else spotted the spike. You already know what. You want why.
The cluster examined as a group: what the calls had in common, which side returned the 403, and whether it is still happening.
A threshold search rather than a failure search. Not every investigation is about something broken.
Every matching call over the duration you named, listed with the detail you would need to open any of them.
Somebody wants to know whether traffic to a destination is growing, falling or steady.
A per-day breakdown. Say “retry for the last 7 days” and it re-runs the same question on the wider window.
How to ask so the numbers come back right
One number and maybe a time. Butler finds the call and says whose side ended it.
What’s happening across my network?
Counts, groupings, thresholds and KPIs across a population of calls.
Routing tables, regex and SDP profiles in plain English.
Ask in French, Spanish or Portuguese, get the answer from the English documentation.
Changes, reports, files and email, each one waiting for your yes.
Paste the complaint in the customer’s own words and let Butler find the call.
Ask Butler what your network did yesterday
Deployed in 48 hours, month to month, in the chat app your team already has open.
By submitting this form, your information will be processed in accordance with our Privacy Policy.