Proactive Alert Monitoring (PAM) relies on machine learning to understand your voice traffic patterns over time and generates an alert when an abnormal voice traffic behavior is detected. Please note that this doesn’t necessarily indicate a network outage or event, but rather serves as an early warning of a potential issue to prompt your support/NOC team to investigate changes in call behavior and perform diagnostics on the internal network.
Changes in calling patterns may be influenced by various factors, such as holidays, national events, or new calling campaigns. This also includes scheduled maintenance that may affect downstream vendors. While Bandwidth can certainly work with you to fine-tune the algorithm, false positives can occur, particularly with low-volume routes and countries.
Glossary
Answer Seizure Ratio (ASR) – total count of 200 SIP response codes (completed calls) / total number of attempted calls
Network Efficiency Ratio (NER) – total count of 200, 404, 486, and 487 SIP response codes / total number of attempted calls
Minutes of Use (MOU) – total minutes of use associated with completed calls
Time Range – time/date window of an event; all times are in Coordinated Universal Time (UTC)
Volume of Attempted Calls – total number of attempted calls in the given time window
Metrics and definitions
Metric | Definition |
Low ASR | The ASR % detected in the given time window is lower than expected. |
Low NER | The NER % detected in the given time window is lower than expected. |
High Volume of Incomplete Calls | The volume of calls that didn’t complete in the given time window is higher than expected. |
Low MOU | The overall MOU in the given time window is lower than expected. |
Low Volume of Completed Calls | The number of calls completed in the given time window is lower than expected. |
Low Volume of Attempted Calls | The number of attempted calls placed in the given time window is lower than expected. |
