IP Library › Granted Patent US 12,639,157
Granted Patent B2
US 12,639,157 · App. 18/790,107 · Granted May 26, 2026

Self-healing latency issues in a cluster network using golden signal telemetry datasets

Inventors: Supriya Kannery (Bengaluru, IN); Rajat Badola (Bengaluru, IN); Philip Shilane (Newtown, PA)
Assignee: Dell Products L.P.
G06F11/0793G06F11/0709G06F11/079
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,639,157
App. No.
18/790,107
Granted
May 26, 2026
Kind
B2
Abstract

A telemetry processing system in a Kubernetes-based cluster network receives golden signal telemetry data from a plurality of pods and formats it into a structured format for validation and storage in a datastore. The golden signals include are generated during data protection operations and are used to reveal problems encountered by the pods. A latency golden signal measures a time between operation request and initiation. A threshold value is defined for the golden signal, and a latency probe is deployed in each pod to monitor the golden signal to detect when an operation exceeds the threshold value indicating a problem with an operation or data object. A self-healing process is called by the probe to initiate remedial measures to address the problem.

Claims (37)

1 . A method of providing a self-healing Kubernetes-based cluster network having a plurality of nodes and pods by processing golden signal telemetry data, comprising:

first registering, with a hardware-based telemetry transmitter, a golden signal comprising telemetry data related to one of traffic, latency, errors, and saturation in a storage system of the network by naming a registering pod, a probe name, an endpoint, and a threshold value;

second registering, with a self-healing service, a hardware-implemented pod-based probe for the golden signal with the telemetry transmitter;

monitoring, by the self-healing service, an activity related to the golden signal in comparison with a respective threshold value; and

calling the probe when an activity value exceeds the respective threshold value as an indication of a problem condition to trigger implementation of remedial steps to address the problem condition and implement self-healing of the cluster network.

2 . The method of claim 1 further comprising:

deploying, within the probe, a debugger in the registering pod; and

performing a debugging operation in the debugger to identify the problem condition.

3 . The method of claim 2 wherein the golden signal comprises a latency golden signal for a data protection operation executed by the cluster network, and wherein the problem condition comprises a program call that is taking an excessive amount of time related to a normal condition, and further wherein the probe comprises a latency probe, and yet further wherein the latency golden signal measures a time between a moment a data protection operation is requested and when the data protection operation is initiated.

4 . The method of claim 3 wherein the cluster network comprises a Santorini network processing containerized data utilizing Kubernetes containerization mechanisms, and wherein the registering pod performs network functions and generates the golden signal as telemetry data for transmission to the self-healing service while performing data protection operations of a backup server.

5 . The method of claim 4 further comprising inputting the telemetry data to a datastore through a telemetry pipeline as input telemetry data, and wherein the telemetry pipeline implements an Open Telemetry (OTEL) protocol, and comprises a collector receiving the input telemetry data through a remote procedure call (RPC) process.

6 . The method of claim 2 further comprising:

collecting, by the debugger, details of the problem condition; and

analyzing time profile logs to identify a program call that is taking the excessive amount of time.

7 . The method of claim 6 further comprising, after the analyzing step, calling the self-healing service to perform the remedial steps to address the problem condition.

8 . The method of claim 7 wherein the problem condition comprises a continuous lock condition of a database, and wherein the remedial steps comprise:

identifying processes that lock the database;

taking a backup of the database;

restarting processes locking the database; and

re-accessing the database to ensure the problem condition is removed.

9 . The method of claim 8 wherein the data protection operation comprises one of a data backup, a data restore, or a data migration operation initiated by a deduplication backup program executed by a backup server.

10 . The method of claim 9 wherein metric datasets of the golden signal are formatted into a structured format for storage in a central datastore accessed by the telemetry transmitter, and wherein the second registering step comprises comparing a format of the golden signal received by the registering pod against a defined structured format.

11 . The method of claim 10 wherein the self-healing service is validated as subscribed to receive golden signal through a subscription process comprising:

implementing a selected transport mechanism through a Representational State Transfer Application Programming Interface (REST API); and

providing the REST API for each transport mechanism to enable subscription by the self-healing service.

12 . A method of debugging issues in a cluster network executing a backup operation processing data from a plurality of pods each periodically generating golden signal metric datasets for transmission to consumers, comprising:

registering, with a hardware-based telemetry transmitter, a latency golden signal measuring a time between a moment a data protection operation is requested and when the data protection operation is initiated, by storing a registering pod identifier, a latency probe identifier, an endpoint, and a threshold value;

detecting, by a hardware-implemented pod-based latency probe, when the threshold value is exceeded by an operation generating the golden signal indicating a problem with the operation; and

calling a self-healing process to identify and trigger implementation of remedial steps to address the problem and implement self-healing of the cluster network, and debug the problem with the operation.

13 . The method of claim 12 wherein the cluster network comprises a Santorini network processing containerized data utilizing Kubernetes containerization mechanisms, and wherein a pod performs network functions and generates the golden signal as telemetry data for transmission to the self-healing service while performing data protection operations of a backup server.

14 . The method of claim 13 wherein the data protection operations comprise at least one of a data backup, data restore, or data migration operation initiated by a deduplication backup program executed by the backup server, and wherein the problem comprises a program call that is taking an excessive amount of time related to a normal condition.

15 . The method of claim 14 further comprising, after an analyzing step, calling the self-healing service to perform remedial steps to address the problem condition.

16 . The method of claim 15 wherein the problem condition comprises a continuous lock condition of a database, and wherein the remedial steps comprise:

identifying processes that lock the database;

taking a backup of the database;

restarting processes locking the database; and

re-accessing the database to ensure the problem condition is removed.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jul 31, 2024
From: KANNERY, SUPRIYA; BADOLA, RAJAT; SHILANE, PHILIP
To: DELL PRODUCTS L.P.
Reel/Frame 068136/0961 →
Continuity (1)
Related Publication 20260037362A1 · Feb 5, 2026
References Cited (30)
US 11411811B2 · Paradkar · 2022 [cited by examiner]
US 12413491B2 · Vegas · 2025 [cited by applicant]
US 20080034003A1 · Stakutis · 2008 [cited by applicant]
US 20080301783A1 · Abrutyn · 2008 [cited by applicant]
US 20170060897A1 · Madaan · 2017 [cited by applicant]
US 20170155622A1 · Vazquez · 2017 [cited by applicant]
US 20190028552A1 · Johnson · 2019 [cited by applicant]
US 20200348973A1 · Kutch · 2020 [cited by applicant]
US 20210006613A1 · Voss-Wolff · 2021 [cited by applicant]
US 20210168150A1 · Ross · 2021 [cited by applicant]
US 20210306899A1 · Mishra · 2021 [cited by applicant]
US 20210342193A1 · Anand · 2021 [cited by applicant]
US 20220147542A1 · Asgar · 2022 [cited by examiner]
US 20220414026A1 · Iyer · 2022 [cited by applicant]
US 20230217273A1 · Viswambharan · 2023 [cited by applicant]
US 20230222043A1 · Lee · 2023 [cited by applicant]
US 20230236922A1 · Nagar · 2023 [cited by applicant]
US 20230244475A1 · Holowaty · 2023 [cited by applicant]
US 20230300218A1 · Karthikeyan · 2023 [cited by applicant]
US 20230362236A1 · Nair · 2023 [cited by applicant]
US 20230393547A1 · Leroy · 2023 [cited by applicant]
US 20230409369A1 · Liu · 2023 [cited by applicant]
US 20240214294A1 · Miriyala · 2024 [cited by applicant]
US 20240223454A1 · Miriyala · 2024 [cited by applicant]
US 20240235970A1 · Sandhaus · 2024 [cited by applicant]
US 20240354290A1 · Scott · 2024 [cited by applicant]
US 20240403137A1 · Rao · 2024 [cited by applicant]
US 20240406277A1 · Henkel · 2024 [cited by applicant]
US 20240421979A1 · Kenchaiah · 2024 [cited by applicant]
US 20250219894A1 · Gupta · 2025 [cited by applicant]