# US 2021/0334186 A1 · Diagnosing and mitigating memory leak in computing nodes

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0334186 A1
- Document ID: US20210334186A1
- Status: Published application
- Kind: A1
- Application number: 16/855,949
- Filed: 2020-04-22
- Published: 2021-10-28
- Earliest claimed: 2020-04-22
- Assignees: Microsoft Technology Licensing, LLC
- Applicants: Microsoft Technology Licensing, LLC
- Inventors: Cong CHEN; Xinsheng YANG; Yingnong DANG; Si QIN
- Web page: https://patentgrep.com/patent/US20210334186A1

## CPC

- H04L 43/0876: Network utilisation, e.g. volume of load or congestion level
- G06F 11/3037: {where the computing system component is a memory, e.g. virtual memory, cache (accessing, addressing or allocating within memory systems or architectures G06F12/00; checking stores for correct operation G11C29/00)}
- G06F 11/3447: {Performance evaluation by modeling}
- G06F 11/3452: {Performance evaluation by statistical analysis}
- G06F 11/3466: {Performance evaluation by tracing or monitoring}
- G06F 12/0253: {Garbage collection, i.e. reclamation of unreferenced memory}
- H04L 41/145: {involving simulating, designing, planning or modelling of a network}
- H04L 41/40: using virtualisation of network functions or resources, e.g. SDN or NFV entities
- H04L 43/20: the monitoring system or the monitored elements being virtualised, abstracted or software-defined entities, e.g. SDN or NFV
- H04L 67/10: in which an application is distributed across nodes in the network (software deployment G06F8/60; multiprogramming arrangements G06F9/46)
- G06F 2201/81: Threshold
- H04L 41/046: comprising network management agents or mobile agents therefor
- H04L 41/0816: the condition being an adaptation, e.g. in response to network events
- H04L 41/0895: Configuration of virtualised networks or elements, e.g. virtualised network function or OpenFlow elements
- H04L 41/0897: by horizontal or vertical scaling of resources, or by migrating entities, e.g. virtual resources or entities
- H04L 43/022: by sampling
- H04L 43/106: using time related information in packets, e.g. by adding timestamps
- H04L 43/16: Threshold monitoring
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

The present disclosure relates to systems, methods, and computer readable media for diagnosing and mitigating memory impact events, such as memory leaks, high memory usage, or other memory issues causing a host node from performing as expected on a cloud computing system. The systems described herein involve receiving locally generated memory usage data from a plurality of host nodes. The systems described herein may aggregate the memory usage data and determine a memory impact diagnosis based on a subset of the aggregated memory usage data. The systems described herein may further apply a mitigation model for mitigating the memory impact event. The systems described herein provide an end-to-end solution for diagnosing and mitigating a variety of memory issues using a dynamic and scalable system that reduces a negative impact of memory leaks and other memory issues on a cloud computing system.

## Claims

### Claim 1

A method, comprising:
collecting memory usage data for a plurality of host nodes on a cloud computing system, the memory usage data including sampled data corresponding to one or more types of memory usage on the plurality of host nodes;
aggregating the memory usage data over one or more predetermined intervals to determine a subset of host nodes from the plurality of host nodes predicted to have memory leaks based on memory usage data for the subset of host nodes satisfying one or more impact metrics;
diagnosing one or more memory leaks for the subset of host nodes based on memory usage data corresponding to the subset of host nodes; and
applying a mitigation model to the portion of the aggregated data to determine and generate a mitigation command to provide to one or more host nodes from the subset of host nodes.

### Claim 2 (depends on 1)

The method of claim 1, wherein collecting memory usage data comprises receiving, from each host node of the plurality of host nodes, locally generated memory usage data, the locally generated memory usage data comprising memory usage data sampled by one or more agents on each of the plurality of host nodes.

### Claim 3 (depends on 2)

The method of claim 2, wherein the one or more agents comprise a first agent and a second agent for collecting different types of memory usage data, wherein the first agent collects memory usage data for one or more application processes, and wherein the second agent collects memory usage data for an operating system and drivers.

### Claim 4 (depends on 1)

The method of claim 1, wherein aggregating the memory usage data comprises determining, for each host node, a relevant time interval, wherein determining the relevant time interval comprises:
determining whether a trend of the memory usage data for a host node has increased over a threshold period of time; and
iteratively modifying the time interval based on determining whether the trend of the memory usage data for the host node has increased over the threshold period of time.

### Claim 5 (depends on 4)

The method of claim 4, wherein aggregating the memory usage data further comprises selectively identifying the portion of the aggregated memory usage data based on timestamps for the portion of the aggregated memory usage data corresponding to the modified time interval.

### Claim 6 (depends on 1)

The method of claim 1, wherein the one or more impact metrics includes one or more of:
a threshold increase in memory usage over a relevant time interval;
a threshold increase in memory usage over a short duration of time;
a threshold decrease in memory usage over the relevant time interval; or
a lack of change in memory usage over the relevant time interval.

### Claim 7 (depends on 1)

The method of claim 1, further comprising:
causing the plurality of host nodes to provide additional memory usage data to an analysis service, the analysis service including a third-party service on the cloud computing system or a different cloud computing system; and
receiving, from the analysis service, information associated with specific host nodes and processes on which memory leaks have occurred.

### Claim 8 (depends on 7)

The method of claim 7, further comprising issuing a rollout stop signal causing a rollout of one or more applications on the cloud computing system to stop.

### Claim 9 (depends on 8)

The method of claim 8, wherein determining the subset of host nodes is further based on an identification of host nodes from the plurality of host nodes affected by the rollout of the one or more applications prior to issuing the rollout stop signal.

### Claim 10 (depends on 1)

The method of claim 1, wherein diagnosing the one or more memory leaks comprises determining a severity of the one or more memory leaks, and wherein applying the mitigation model comprises determining the mitigation action from a plurality of mitigation actions based on the severity of the one or more memory leaks.

### Claim 11 (depends on 1)

The method of claim 1, wherein diagnosing the one or more memory leaks comprises determining a type of the one or more memory leaks, and wherein applying the mitigation model comprises determining the mitigation action from the plurality of mitigation actions based on the determined type of the one or more memory leaks.

### Claim 12 (depends on 1)

The method of claim 1, wherein diagnosing the one or more memory leaks comprises determining a type of the one or more memory leaks and a severity of the one or more memory leaks, and wherein applying the mitigation model comprises iterating through a mitigation tree including a plurality of mitigation commands to identify the mitigation command from the plurality of mitigation commands to provide to the one or more host nodes from the subset of host nodes.

### Claim 13

A system, comprising:
one or more processors;
memory in electronic communication with the one or more processors; and
instructions stored in the memory, the instructions being executable by the one or more processors to:
collect memory usage data for a plurality of host nodes on a cloud computing system, the memory usage data including sampled data corresponding to one or more types of memory usage on the plurality of host nodes;
aggregate the memory usage data over one or more predetermined intervals to determine a subset of host nodes from the plurality of host nodes predicted to have memory leaks based on memory usage data for the subset of host nodes satisfying one or more impact metrics;
diagnose one or more memory leaks for the subset of host nodes based on memory usage data corresponding to the subset of host nodes; and
apply a mitigation model to the portion of the aggregated data to determine and generate a mitigation command to provide to one or more host nodes from the subset of host nodes.

### Claim 14 (depends on 13)

The system of claim 13,
wherein collecting memory usage data comprises receiving, from each host node of the plurality of host nodes, locally generated memory usage data, the locally generated memory usage data comprising memory usage data sampled by one or more agents on each of the plurality of host nodes, and
wherein the one or more agents comprise a first agent and a second agent for collecting different types of memory usage data, wherein the first agent collects memory usage data for one or more application processes, and wherein the second agent collects memory usage data for an operating system and drivers.

### Claim 15 (depends on 13)

The system of claim 13, wherein aggregating the memory usage data comprises determining, for each host node, a relevant time interval, wherein determining the relevant time interval comprises:
determining whether a trend of the memory usage data for a host node has increased over a threshold period of time; and
iteratively modifying the time interval based on determining whether the trend of the memory usage data for the host node has increased over the threshold period of time.

### Claim 16 (depends on 13)

The system of claim 13, further comprising instructions being executable by the one or more processors to:
cause the plurality of host nodes to provide additional memory usage data to an analysis service, the analysis service including a third-party service on the cloud computing system or a different cloud computing system;
receive, from the analysis service, information associated with specific host nodes and processes on which memory leaks have occurred; and
issue a rollout stop signal causing a rollout of one or more applications on the cloud computing system to stop, wherein determining the subset of host nodes is further based on an identification of host nodes from the plurality of host nodes affected by the rollout of the one or more applications prior to issuing the rollout stop signal.

### Claim 17

A non-transitory computer readable medium storing instructions thereon that, when executed by one or more processors, causes one or more server devices to:
collect memory usage data for a plurality of host nodes on a cloud computing system, the memory usage data including sampled data corresponding to one or more types of memory usage on the plurality of host nodes;
aggregate the memory usage data over one or more predetermined intervals to determine a subset of host nodes from the plurality of host nodes predicted to have memory leaks based on memory usage data for the subset of host nodes satisfying one or more impact metrics;
diagnose one or more memory leaks for the subset of host nodes based on memory usage data corresponding to the subset of host nodes; and
apply a mitigation model to the portion of the aggregated data to determine and generate a mitigation command to provide to one or more host nodes from the subset of host nodes.

### Claim 18 (depends on 17)

The non-transitory computer readable medium of claim 17,
wherein collecting memory usage data comprises receiving, from each host node of the plurality of host nodes, locally generated memory usage data, the locally generated memory usage data comprising memory usage data sampled by one or more agents on each of the plurality of host nodes, and
wherein the one or more agents comprise a first agent and a second agent for collecting different types of memory usage data, wherein the first agent collects memory usage data for one or more application processes, and wherein the second agent collects memory usage data for an operating system and drivers.

### Claim 19 (depends on 17)

The non-transitory computer readable medium of claim 17, wherein aggregating the memory usage data comprises determining, for each host node, a relevant time interval, wherein determining the relevant time interval comprises:
determining whether a trend of the memory usage data for a host node has increased over a threshold period of time; and
iteratively modifying the time interval based on determining whether the trend of the memory usage data for the host node has increased over the threshold period of time.

### Claim 20 (depends on 17)

The non-transitory computer readable medium of claim 17, further comprising instructions that, when executed by the one or more processors, causes one or more server devices to:
cause the plurality of host nodes to provide additional memory usage data to an analysis service, the analysis service including a third-party service on the cloud computing system or a different cloud computing system;
receive, from the analysis service, information associated with specific host nodes and processes on which memory leaks have occurred; and
issue a rollout stop signal causing a rollout of one or more applications on the cloud computing system to stop, wherein determining the subset of host nodes is further based on an identification of host nodes from the plurality of host nodes affected by the rollout of the one or more applications prior to issuing the rollout stop signal.
