Amazon Technologies, Inc.

Lawrence Shao-Shien Young

Christopher Thomas

Koji Hashimoto

Pratheek Benamanahalli Nagaraj

A web services system includes a plurality of computing devices configurable to implement a plurality of compute instances of a provider network. One or more of the computing devices is configured to implement a network device and a resource manager. The network device is configured to receive traffic directed to an application. The resource manager is configured to determine an amount of computing resources employed by the network device to process a first workload presented to the network device over a predetermined precedent time interval, to determine, based on the amount of computing resources employed by the network device, an amount of computing resources needed by the network device to process a second workload; and to allocate the amount of computing resources to the network device.

Controlling network device capacity

A system includes multiple computing nodes. Each node includes a processor and memory. The computing nodes comprise a plurality of instances of an application, a networking device and a health service. The networking device is configured to receive requests directed to the application instances, to distribute the requests across the application instances, and, for each request, to determine a status code and an elapsed time value and store the status codes and elapsed time values for the request in a log. The health service is configured to provide error rate and average or percentile latency metrics based on request status codes and elapsed time values from the log. The metrics comprise a first set of error rate and average or percentile latency metrics associated with the networking device and a second set of error rate and average or percentile latency metrics associated with the application instances.

Automated method and system for diagnosing load performance issues

Disclosed are techniques for determining possible causes of anomalous state in computing nodes. A computing node is analyzed for a given time period against another computing node over another time period. Individual metrics such as network bandwidth are measured and a subset of these metrics that are determined to be most likely related to the cause of anomalous state are provided to a user.

Automated anomaly detection

Computing resource service providers allow customers to execute computer systems on hardware provided by the computing resource service provider. The hardware may be distributed between various geographic locations connected by a network. A load balancer may be provided to distribute traffic between the computer systems. Furthermore, computing resource service provider may cause computing resources to be allocated or deallocated to the load balancer based at least in part on various attributes of the computer systems the load balancer is responsible for distributing traffic to. The various attributes may include a capacity of the computer systems.

Predictive load balancer resource management

A web services system includes a plurality of computing nodes, each comprising a processor and memory coupled to the at least one processor. The computing nodes include a network device and a management service. The network device is configured to receive service requests directed to an application. The management service allocates resources to the network device, and is configured to measure usage of a plurality of different types of the resources by the network device, and to determine a value of resource utilization for the network device. The value is expressed according to a standard that represents all the different types of the resources, and the value is selected based on one of the different types of the resources that is most highly utilized by the network device.

Reserved network device capacity

A system for mitigating operational impairment of a computer system includes a plurality of computing nodes. Each of the computing nodes includes a processor and memory coupled to the processor. The computing nodes are configured to: implement a plurality of computing devices; identify an impairment in operation of a first of the computing devices; execute a first remediation operation to mitigate the impairment; analyze operation of the first of the computing devices after execution of the first remediation operation to determine whether the first remediation operation is effective to mitigate the impairment; and change a frequency with which the first remediation operation is applied to mitigate the impairment based on whether the first remediation operation is deemed effective to mitigate the impairment.

Growing community of inventors

Mercer Island, WA, United States of America

Lawrence Shao-Shien Young