Assurance of policy based alerting
Embodiments provide for assuring policy based alerting, via clustering, via a first neural network, operational data reported from a network into a plurality of anomalies organized into several clusters; correlating, via the first neural network, alerts received from devices in the network according to the several clusters; determining, via the second neural network, anomaly impacts in the several clusters from the filtered alerts; in response to determining that the anomaly impacts for a first cluster exceed an alerting threshold: identifying a first shared node in the first cluster; identifying a second cluster including a second shared node matching the first shared node that has not been determined to exceed the alerting threshold; and transmitting an alert for the first cluster and the second cluster; and in response to receiving a response to the alert, updating, via the second neural network, the first neural network.
1. A method, comprising:
training a first neural network to cluster devices into geographic clusters based on physical location of the devices;
training a second neural network, distinct from the first neural network, to cluster the devices into topological clusters based on connections between the devices in a network;
training a third neural network, distinct from the first and second neural networks, to cluster devices into call flow clusters based on device type, wherein the first neural network, the second neural network, and the third neural network form an anomaly filtering engine;
clustering, in real-time, using the anomaly filtering engine, devices in the network based on operational data reported from the network into a set of geographic clusters, a set topological clusters distinct from the set of geographic clusters, and a set of call flow clusters distinct from the set of geographic clusters and the set of topological clusters;
correlating, in real-time, via the anomaly filtering engine, alerts received from devices in the network according to the set of geographic clusters, the set of topological clusters, and the set of call flow clusters;
detecting, in real-time, via a fourth neural network, anomaly impacts in the set of geographic clusters, the set of topological clusters, and the set of call flow clusters from the alerts, wherein the anomaly impacts comprise network conditions affecting an operation of the network;
in response to determining that the anomaly impacts for a first geographic cluster, of the set of geographic clusters, exceed an alerting threshold:
identifying, in real-time, a first device of a first plurality of devices in the first geographic cluster, wherein the first device corresponds to a lowest-tiered device in a network topology of the network where communications pathways of the first plurality of devices in the first geographic cluster pass through the lowest-tiered device;
identifying, in real-time, a first call flow cluster, of the set of call flow clusters, that includes the first device, wherein the anomaly impacts for the first call flow cluster have not been determined to exceed the alerting threshold, wherein the first call flow cluster comprises a second plurality of devices including the first device, wherein the second plurality of devices comprises at least one device outside of the first plurality of devices, and wherein communications pathways of the second plurality of devices in the first call flow cluster pass through the lowest-tiered device; and
transmitting, in real-time, an alert indicating the anomaly impacts for both the first geographic cluster and the first call flow cluster; and
in response to receiving a response to the alert, updating, in real-time, via the fourth neural network, the anomaly filtering engine wherein the updates to anomaly filtering engine adjust how the first neural network, the second neural network, and the third neural network identify clusters in the operational data.
2. The method of claim 1 , wherein the operational data is gathered via data pipelines defined in the network for associated Key Performance Indicators that store the operational data in a data lake accessible to the first neural network for a predetermined length of time.
3. The method of claim 1 , wherein a device represented by a first shared node in the first geographic cluster is also included as a non- shared node in a second geographic cluster, wherein the alert does not include the second geographic cluster.
4. The method of claim 1 , further comprising:
prior to clustering the operational data:
normalizing the operational data by removing outliers from the operational data; and
smoothing the operational data by adding predicted data points to represent data points missing from the operational data.
5. The method of claim 1 , wherein the second neural network generates the response to the alert without user input, further comprising:
transmitting the response to at least one device included in the first geographic cluster.
6. The method of claim 1 , wherein the alerting threshold is dynamically adjusted based on a number of end users affected by a given alert.
7. A system, comprising:
a processor; and
a memory, including instructions that when performed by the processor enable the processor to:
train a first neural network to cluster devices into geographic clusters based on physical location of the devices;
train a second neural network, distinct from the first neural network, to cluster the devices into topological clusters based on connections between the devices in a network;
train a third neural network, distinct from the first and second neural networks, to cluster devices into call flow clusters based on device type, wherein the first neural network, the second neural network, and the third neural network form an anomaly filtering engine;
cluster, in real-time, via using the anomaly filtering engine, devices in the network based on operational data reported from the network into a set of geographic clusters, a set topological clusters distinct from the set of geographic clusters, and a set of call flow clusters distinct from the set of geographic clusters and the set of topological clusters;
correlate, in real-time, via the anomaly filtering engine, alerts received from devices in the network according to the set of geographic clusters, the set of topological clusters, and the set of call flow clusters;
detect, in real-time, via a fourth neural network, anomaly impacts in the set of geographic clusters, the set of topological clusters, and the set of call flow clusters from the alerts, wherein the anomaly impacts comprise network conditions affecting an operation of the network;
in response to determining that the anomaly impacts for a first geographic cluster, of the set of geographic clusters, exceed an alerting threshold:
identify, in real-time, a first device of a first plurality of devices in the first geographic cluster, wherein the first device corresponds to a lowest-tiered device in a network topology of the network where communications pathways of the first plurality of devices in the first geographic cluster pass through the lowest-tiered device;
identify, in real-time, a first call flow cluster, of the set of call flow clusters, that includes the first device, wherein the anomaly impacts for the first call flow cluster have not been determined to exceed the alerting threshold, wherein the first call flow cluster comprises a second plurality of devices including the first device, wherein the second plurality of devices comprises at least one device outside of the first plurality of devices, and wherein communications pathways of the second plurality of devices in the first call flow cluster pass through the lowest-tiered device; and
transmit, in real-time, an alert indicating both the first geographic cluster and the first call flow cluster; and
in response to receiving a response to the alert, update, in real-time, via the fourth neural network, the anomaly filtering engine wherein the updates to anomaly filtering engine adjust how the first neural network, the second neural network, and the third neural network identify clusters in the operational data.
8. The system of claim 7 , wherein the operational data is gathered via data pipelines defined in the network for associated Key Performance Indicators that store the operational data in a data lake accessible to the first neural network for a predetermined length of time.
9. The system of claim 7 , wherein a device represented by a first shared node of the first geographic cluster is also included as a non- shared node in a second geographic cluster, wherein the alert does not include the second geographic cluster.
10. The system of claim 7 , further comprising:
prior to clustering the operational data, the instructions further enable the processor to:
normalize the operational data by removing outliers from the operational data; and
smooth the operational data by adding predicted data points to represent data points missing from the operational data.
11. The system of claim 7 , wherein the second neural network generates the response to the alert without user input, and the instructions further enable the processor to:
transmit the response to at least one device included in the first geographic cluster.
12. The system of claim 7 , wherein the instructions further enable the processor is to dynamically adjusted the alerting threshold based on a number of end users affected by a given alert.
13. A non-transitory computer readable medium including processor executable instruction for performing an operation, comprising:
training a first neural network to cluster devices into geographic clusters based on physical location of the devices;
training a second neural network, distinct from the first neural network, to cluster the devices into topological clusters based on connections between the devices in a network;
training a third neural network, distinct from the first and second neural networks, to cluster devices into call flow clusters based on device type, wherein the first neural network, the second neural network, and the third neural network form an anomaly filtering engine;
clustering, in real-time, using the anomaly filtering engine, devices in the network based on operational data reported from the network into a set of geographic clusters, a set topological clusters distinct from the set of geographic clusters, and a set of call flow clusters distinct from the set of geographic clusters and the set of topological clusters;
correlating, in real-time, via the anomaly filtering engine, alerts received from devices in the network according to the set of geographic clusters, the set of topological clusters, and the set of call flow clusters;
detecting, in real-time, via a fourth neural network, anomaly impacts in the set of geographic clusters, the set of topological clusters, and the set of call flow clusters from the alerts, wherein the anomaly impacts comprise network conditions affecting an operation of the network;
in response to determining that the anomaly impacts for a first geographic cluster, of the set of geographic clusters, exceed an alerting threshold:
identifying, in real-time, a first device of a first plurality of devices in the first geographic cluster, wherein the first device corresponds to a lowest-tiered device in a network topology of the network where communications pathways of the first plurality of devices in the first geographic cluster pass through the lowest-tiered device;
identifying, in real-time, a first call flow cluster, of the set of call flow clusters, that includes the first device, wherein the anomaly impacts for the first call flow cluster have not been determined to exceed the alerting threshold, wherein the first call flow cluster comprises a second plurality of devices including the first device, wherein the second plurality of devices comprises at least one device outside of the first plurality of devices, and wherein communications pathways of the second plurality of devices in the first call flow cluster pass through the lowest-tiered device; and
transmitting, in real-time, an alert indicating both the first geographic cluster and the first call flow cluster; and
in response to receiving a response to the alert, updating, in real-time, via the fourth neural network, the anomaly filtering engine wherein the updates to anomaly filtering engine adjust how the first neural network, the second neural network, and the third neural network identify clusters in the operational data.
14. The non-transitory computer readable medium of claim 13 , wherein the operational data is gathered via data pipelines defined in the network for associated Key Performance Indicators that store the operational data in a data lake accessible to the first neural network for a predetermined length of time.
15. The non-transitory computer readable medium of claim 13 , wherein a device represented by a first shared node of the first geographic cluster is also included as a non-shared node in a second geographic cluster, wherein the alert does not include the second geographic cluster.
16. The non-transitory computer readable medium of claim 13 , further comprising:
prior to clustering the operational data:
normalizing the operational data by removing outliers from the operational data; and
smoothing the operational data by adding predicted data points to represent data points missing from the operational data.
17. The non-transitory computer readable medium of claim 13 , wherein the second neural network generates the response to the alert without user input, further comprising:
transmitting the response to at least one device included in the first geographic cluster.
18. The non-transitory computer readable medium of claim 13 , wherein the alerting threshold is dynamically adjusted based on a number of end users affected by a given alert.
19. The method of claim 1 , further comprising in response to receiving the alert, adjusting the first plurality of devices to at least one of: shift loads in the network, deactivate one of the first plurality of devices, activate one of the first plurality of devices, move one of the first plurality of devices to or from a standby mode, update software running on one of the first plurality of devices, or quarantine one of the first plurality of devices within the network.
20. The system of claim 7 , further comprising in response to receiving the alert, adjusting the first plurality of devices to at least one of: shift loads in the network, deactivate one of the first plurality of devices, activate one of the first plurality of devices, move one of the first plurality of devices to or from a standby mode, update software running on one of the first plurality of devices, or quarantine one of the first plurality of devices within the network.