Automated root cause analysis of underperforming video streams by using language transformers on support ticket systems

1. A method of correcting a content delivery infrastructure, the method comprising:

receiving a request to resolve reported issues for the content delivery infrastructure;

collecting content delivery metrics for the content delivery infrastructure, in response to receiving the request, the content delivery metrics measuring operations of delivering streaming content in the content delivery infrastructure;

executing a machine learning model in one or more cloud regions that process the streaming content within the content delivery infrastructure prior to the streaming content reaching a user device, based on input of the request and the content delivery metrics to generate a set of possible resolutions to the reported issues with confidence ratings, wherein executing the machine learning model comprises converting text within the content delivery metrics into numerical values as input to the machine learning model, and wherein the machine learning model is a transformer machine learning model adapted to the content delivery infrastructure based on the content delivery metrics for the content delivery infrastructure; and

implementing an automated solution to improve the operations of delivering the streaming content in the content delivery infrastructure based on a resolution from the set of possible resolutions, in response to the resolution having a confidence rating above a threshold.

2. The method of claim 1 , further comprising:

generating a synthesized resolution or generating a link to an existing resolution.

3. The method of claim 1 , further comprising:

providing the resolution to a user of a service provider in the content delivery infrastructure in response to the resolution having a confidence rating above the threshold that does not have the automated solution.

4. The method of claim 1 , further comprising:

providing the set of possible resolutions to a user of a service provider in the content delivery infrastructure in response to the set of possible resolutions not having a confidence rating above the threshold.

5. The method of claim 1 , further comprising:

receiving a set of resolved incidents for a reference content delivery infrastructure; and

retraining adversarial machine learning models for a target content delivery infrastructure.

6. The method of claim 5 , further comprising:

determining whether a machine learning model for the target content delivery infrastructure generates a high probability score in response to an input with a known resolution.

7. The method of claim 6 , further comprising:

receiving video quality assessment, source reliability score, and video stability assessment aligned with the set of resolved incidents and geographic locations of components of the content delivery infrastructure.

8. The method of claim 1 , further comprising:

generating a summary of an automated solution that is sent to a user through an issue tracking service;

receiving feedback on the automated solution from the user; and

retraining the machine learning model with the feedback.

9. A non-transitory machine-readable storage medium that provides instructions that, if executed by a processor, will cause the processor to perform operations of a method of correcting a content delivery infrastructure, the operations comprising:

receiving a request to resolve reported issues for the content delivery infrastructure;

collecting content delivery metrics for the content delivery infrastructure, in response to receiving the request, the content delivery metrics measuring operations of delivering streaming content in the content delivery infrastructure;

executing a machine learning model in one or more cloud regions that process the streaming content within the content delivery infrastructure prior to the streaming content reaching a user device, based on input of the request and the content delivery metrics to generate a set of possible resolutions to the reported issues with confidence ratings, wherein executing the machine learning model comprises converting text within the content delivery metrics into numerical values as input to the machine learning model, and wherein the machine learning model is a transformer machine learning model adapted to the content delivery infrastructure based on the content delivery metrics for the content delivery infrastructure; and

implementing an automated solution to improve the operations of delivering the streaming content in the content delivery infrastructure based on a resolution from the set of possible resolutions, in response to the resolution having a confidence rating above a threshold.

10. The non-transitory machine-readable storage medium of claim 9 , wherein the operations further comprise:

generating a synthesized resolution or generating a link to an existing resolution.

11. The non-transitory machine-readable storage medium of claim 9 , wherein the operations further comprise:

providing the resolution to a user of a service provider in the content delivery infrastructure in response to the resolution having a confidence rating above the threshold that does not have the automated solution.

12. The non-transitory machine-readable storage medium of claim 9 , wherein the operations further comprise:

providing the set of possible resolutions to a user of a service provider in the content delivery infrastructure in response to the set of possible resolutions not having a confidence rating above the threshold.

13. The non-transitory machine-readable storage medium of claim 9 , wherein the operations further comprise:

receiving a set of resolved incidents for a reference content delivery infrastructure; and

retraining adversarial machine learning models for a target content delivery infrastructure.

14. The non-transitory machine-readable storage medium of claim 13 , wherein the operations further comprise:

determining whether a machine learning model for the target content delivery infrastructure generates a high probability score in response to an input with a known resolution.

15. The non-transitory machine-readable storage medium of claim 14 , wherein the operations further comprise:

receiving video quality assessment, source reliability score, and video stability assessment aligned with the set of resolved incidents and geographic locations of components of the content delivery infrastructure.

16. An electronic device to implement a method of measuring video stream visual stability, the electronic device comprising:

a non-transitory machine-readable storage medium having stored therein a root cause analysis service; and

a processor coupled to the non-transitory machine-readable storage medium, the processor to execute the root cause analysis service, the root cause analysis service to receive a request to resolve reported issues for a content delivery infrastructure, to collect content delivery metrics for the content delivery infrastructure, the content delivery metrics measuring operations of delivering streaming content in the content delivery infrastructure, to execute a machine learning model in one or more cloud regions that process the streaming content within the content delivery infrastructure prior to the streaming content reaching a user device, based on input of the request and the content delivery metrics to generate a set of possible resolutions for the reported issues with confidence ratings, wherein executing the machine learning model comprises converting text within the content delivery metrics into numerical values as input to the machine learning model, and wherein the machine learning model is a transformer machine learning model adapted to the content delivery infrastructure based on the content delivery metrics for the content delivery infrastructure, and to implement an automated solution to improve the operations of delivering the streaming content in the content delivery infrastructure based on a resolution from the set of possible resolutions, in response to the resolution having a confidence rating above a threshold.

17. The electronic device of claim 16 , wherein the root cause analysis service is further to generate a synthesized resolution or generating a link to an existing resolution.

18. The electronic device of claim 16 , wherein the root cause analysis service is further to provide the resolution to a user of a service provider in the content delivery infrastructure in response to the resolution having a confidence rating above the threshold that does not have the automated solution.

19. The electronic device of claim 16 , wherein the root cause analysis service is further to provide the set of possible resolutions to a user of a service provider in the content delivery infrastructure in response to the set of possible resolutions not having a confidence rating above the threshold.

20. The electronic device of claim 16 , wherein the root cause analysis service is further to receive a set of resolved incidents for a reference content delivery infrastructure, and retrain adversarial machine learning models for a target content delivery infrastructure.

21. The electronic device of claim 20 , wherein the root cause analysis service is further to determine whether a machine learning model for the target content delivery infrastructure generates a high probability score in response to an input with a known resolution.

22. The electronic device of claim 21 , wherein the root cause analysis service is further to receive video quality assessment, source reliability score, or video stability assessment aligned with the set of resolved incidents.