Latency Report¶
Generating the Latency Report¶
The Latency Report highlights calls exceeding a defined latency threshold. For example, supervisors can generate a report of all calls where the P90 end-to-end latency is greater than 500 milliseconds (ms).
Note
In this beta version, to generate Latency reports, enable capturing latency metrics by configuring these call control parameters:
notifySttLatency = true
vadEnable = true
vadVendor = silero
To generate a Latency Report:
- On the Reports tab, click New Report.
- In the list of reports, choose the Latency Report.
- Enter the Report Name and Report Description.
- Select a Time Period.
-
Select Data Filters from the following.
P90 Values Threshold Values (ms) P90 End-to-End Latency For example, 200 ms P90 ASR Latency For example, 200 ms P90 Bot Response Time For example, 200 ms P90 TTS Latency For example, 200 ms
Note
Each P90 value can be evaluated using the following conditions:
Greater than or equal to (≥), Less than or equal to (≤), Greater than (>), Less than (<), Equal to (=).
- Select Report Generation and choose from the following options:
- Click Create.
- An instance of the Latency Report is generated.
-
Click the Run icon next to the instance of the report.
While generating reports, you have two options: Run On Demand and Schedule.
- Run On Demand: While creating a new report or editing an existing report, the schedule of the report can be set to run on demand.
- Schedule: While creating a new report or editing an existing report, the schedule of the report can be changed. The use of the schedule is the standard way to generate a report. If the report is scheduled for later, then it will automatically be generated on the scheduled day and time without the need to click Run.
- Click the CSV icon and generate the output of that report, as shown below.
Note
The red dot beside the report indicates that it was generated within the last 24 hours. You can download, save, and view the report in CSV format.
About Report Fields¶
Field | Description | Example |
---|---|---|
Session ID | Unique identifier for each interaction session. | a12b34c56d |
Start Time | Timestamp when the session started. | 2025-09-24 10:56:34 |
Agent | Name of the agent handling the session. | John Doe |
Agent Email | Email address of the agent. | john.doe@example.com |
Queue Name | Name of the queue the interaction was routed through. | Support_Queue |
Customer | Identifier or name of the customer involved in the session. | Jane Doe |
Duration | Total time the session lasted. | 00:12:45 |
P90 ASR Latency (ms) | 90th percentile latency of Automatic Speech Recognition, in milliseconds. Indicates how long speech-to-text processing takes. | 250 |
P90 Bot Latency (ms) | 90th percentile latency for the bot’s response processing, in milliseconds. Shows system responsiveness. | 120 |
P90 TTS Latency (ms) | 90th percentile latency of Text-to-Speech processing, in milliseconds. Represents the time to convert text into audio. | 180 |
P90 End To End Latency (ms) | 90th percentile overall latency from customer speech to agent/bot reply, in milliseconds. Combines ASR, bot, and TTS processing times. | 600 |