The Conviva alerts feature lets you automate the process of monitoring key quality metrics and emailing team members when alert conditions occur. You might watch for increases in Video Start Failures Technical (VSF-T) or Exits Before Video Start (EBVS), both of which indicate viewers are not seeing the content they want. Or, you can notify your programming team when spikes in concurrent plays indicate a sudden increase in viewers.
Note: Conviva generates alerts based on real-time data to minimize reporting latency, while Conviva renders data in Video to ensure data is processed with a high level of confidence. As a result, there may be a delay of up to 5-10 minutes between when an alert is generated and when the related data is rendered for analysis.
Note: Each AI alert cohort requires an on-going minimum of 25 attempts, concurrent plays, or ended plays during each 1-minute interval to qualify for alerting based on the AI alert sensitivity settings. AI alert sensitivity no longer considers deviation as a sensitivity control.
These are the metrics that you can monitor for alerts :
| Monitoring This Metric |
|---|
| Average Peak Bitrate |
| Concurrent Plays |
| Exit Before Video Starts (% of Attempts) |
| Rebuffering Ratio (% of Concurrent Plays) |
| Total Attempts |
| Video Start Failures (% of Attempts) |
| Video Start Failures Technical (% of Attempts related to technical issues) |
| Video Start Failures Business (% of Attempts related to business issues) |
| Video Playback Failures (% of Ended Plays) |
| Video Playback Failures Technical (% of Ended Plays related to technical issues) |
| Video Playback Failures Business (% of Ended Plays related to business issues) |
| Video Startup Time (VST) |
| Video Restart Time (VRT) |
| Connection Induced Rebuffering Ratio |
Note: The availability of concurrent plays is limited. If you require this feature, kindly contact your account manager.
For detailed metric definitions, see the Metric Dictionary.
Metrics Aggregation
With the exception of Attempts, which are summed, and Concurrent Plays, which look at the maximum value, all other Diagnostics metrics are aggregated as weighted averages. Basically, in order to tell when an event occurs, we look at the metric value and the size of the sample. For example, a VSF of 12% and a sample of 100 attempts. If both values are over their respective thresholds, the alert fires. If, however, the metric is high enough but the sample isn't large enough, the systems collect additional sample data, using up to 60 minutes of historical data, until the required sample size and confidence level are reached. At that point, the alert fires with details about the metric, sample size and collected data.
This table shows default sample values for each metric.
| Metric (unit) | Sample | Sample Units |
|---|---|---|
| Average Peak Bitrate (bps) | 500 | Plays |
| Concurrent Plays | 5 |
Minutes Sample size is limited to a maximum of 60 minutes. |
| EBVS (%) | 1000 | Attempts |
| Rebuffering Ratio (%) | 500 | Plays |
| Total Attempts (count) | 5 |
Minutes Sample size is limited to a maximum of 60 minutes. |
| VSF (%) | 1000 | Attempts |
| VSF-T (%) | 1000 | Attempts |
| VSF-B (%) | 1000 | Attempts |
| VST (seconds) | 1000 | Attempts |
| VPF (%) | 1000 | Ended Plays |
| VPF-T (%) | 1000 | Ended Plays |
| VPF-B (%) | 1000 | Ended Plays |
| VRT (%) | 1000 | Ended Plays |
| CIRR (%) | 1000 | Plays |
For more information on Alerts follow one of the following links:
- Alerts
- Creating Alerts
- Managing Alerts
- Advanced Settings
- Viewing Alert Diagnostics
- Viewing Alert Events
- Business Error Configuration
- Automatic Insights