Anomaly detection triggered proactive rerouting for software as a service (SaaS) application traffic

1. A method comprising:

forming, by a device, a telemetry cluster by applying clustering to telemetry data regarding a plurality of network connections over an access network between a plurality of entities and a remote application;

applying, by the device, anomaly detection to the telemetry cluster, to identify one or more anomalies;

making, by the device, a determination as to whether the one or more anomalies are attributable to the remote application or to the access network by determining a temporal correlation between the one or more anomalies and changes to network destinations associated with the remote application; and

initiating, by the device, a corrective measure, based on the determination as to whether the one or more anomalies are attributable to the remote application or to the access network.

2. The method as in claim 1 , wherein forming the telemetry cluster comprises:

selecting the plurality of entities based on the plurality of entities being located in a particular geographic location.

3. The method as in claim 1 , wherein the telemetry data is indicative of one or more of: jitter, loss, or delays.

4. The method as in claim 1 , wherein the device comprises a controller for a plurality of routers.

5. The method as in claim 1 , further comprising:

identifying the changes to the network destinations associated with the remote application using Domain Name System (DNS) resolution data.

6. The method as in claim 1 , wherein making the determination as to whether the one or more anomalies are attributable to the remote application or to the access network comprises:

generating a causal graph based in part on the one or more anomalies; and

analyzing the causal graph, to make the determination.

7. The method as in claim 1 , further comprising:

determining whether the one or more anomalies are seasonal.

8. The method as in claim 7 , wherein, when the one or more anomalies are determined to be seasonal, the corrective measure comprises moving traffic in the access network to different network paths that do not exhibit characteristics of the one or more anomalies.

9. The method as in claim 1 , wherein the remote application is a software as a service (SaaS) application.

10. An apparatus, comprising:

one or more network interfaces;

a processor coupled to the one or more network interfaces and configured to execute one or more processes; and

a memory configured to store a process that is executable by the processor, the process when executed configured to:

form a telemetry cluster by applying clustering to telemetry data regarding a plurality of network connections over an access network between a plurality of entities and remote application;

apply anomaly detection to the telemetry cluster, to identify one or more anomalies;

make a determination as to whether the one or more anomalies are attributable to the remote application or to the access network by determining a temporal correlation between the one or more anomalies and changes to network destinations associated with the remote application; and

initiate a corrective measure, based on the determination as to whether the one or more anomalies are attributable to the remote application or to the access network.

11. The apparatus as in claim 10 , wherein the apparatus forms the telemetry cluster by:

selecting the plurality of entities based on the plurality of entities being located in a particular geographic location.

12. The apparatus as in claim 10 , wherein the telemetry data is indicative of one or more of: jitter, loss, or delays.

13. The apparatus as in claim 10 , wherein the process when executed is further configured to:

identify the changes to the network destinations associated with the remote application using Domain Name System (DNS) resolution data.

14. The apparatus as in claim 10 , wherein the apparatus makes the determination as to whether the one or more anomalies are attributable to the remote application or to the access network by:

generating a causal graph based in part on the one or more anomalies; and

analyzing the causal graph, to make the determination.

15. The apparatus as in claim 10 , wherein the process when executed is further configured to:

determine whether the one or more anomalies are seasonal.

16. The apparatus as in claim 15 , wherein, when the one or more anomalies are determined to be seasonal, the corrective measure comprises moving traffic in the access network to different network paths that do not exhibit characteristics of the one or more anomalies.

17. The apparatus as in claim 10 , wherein the remote application is a software as a service (SaaS) application.

18. A tangible, non-transitory, computer-readable medium storing program instructions that cause a device to execute a process comprising:

forming, by the device, a telemetry cluster by applying clustering to telemetry data regarding a plurality of network connections over an access network between a plurality of entities and a remote application;

applying, by the device, anomaly detection to the telemetry cluster, to identify one or more anomalies;

making, by the device, a determination as to whether the one or more anomalies are attributable to the remote application or to the access network by determining a temporal correlation between the one or more anomalies and changes to network destinations associated with the remote application; and

initiating, by the device, a corrective measure, based on the determination as to whether the one or more anomalies are attributable to the remote application or to the access network.