Abstract: This disclosure describes techniques that include receiving underlay flow data from a network having an underlay network and one or more overlay network, storing information identifying, for each underlay data flow, an overlay network, displaying, within a display, a topological map of at least a portion of the underlay network, highlighting a data path through the displayed topological map, the highlighted data path extending through the underlay network from the underlay network source of the respective underlay data flow to the underlay network destination of the respective underlay data flow; receiving a request for metrics associated with the highlighted data path, wherein receiving a request includes receiving, via a graphical user interface, an indication selecting at least a portion of the highlighted data path; and displaying, proximate to the highlighted data path, metrics associated with data traffic through the selected portion of the highlighted data path.

Abstract: An example control plane that is executed on one or more processors in a distributed computing system is configured to receive an indication of a node to be onboarded into the distributed computing system, wherein the node comprises one of a compute node or a network device node, to discover one or more compute resources or network device resources that are associated with the node, and to assign, based on the discovery, the node to a collector that is executed in the distributed computing system, wherein the collector is configured to collect real-time telemetry data for the node during operation of the node. The control plane is further configured to receive, from the collector, the real-time telemetry data for the node that is collected by the collector, and to output, for display, a visual representation of the real-time telemetry data for the node.

Abstract: An alarm service can receive an alarm rule as an “intent” that defines a rule in a high level “natural language.” An alarm rule compiler can receive the intent and translate the high level intent into one or more lower level rules that can be programmatically processed by multiple alarm rule execution engines. Devices in a network system can be associated with alarm rule execution engines in a distributed manner. For example, devices in a network can be associated with different instances of an alarm rule execution engine, thus distributing the resource usage for obtaining telemetry data and processing alarms with respect to the devices in a network across multiple alarm rule execution engines.

Abstract: The disclosure describes techniques for network monitoring and fault localization. For example, a controller comprises one or more processors operably coupled to a memory configured to: receive a first one or more Quality of Experience (QoE) metrics measured by a first probe traversing a first path comprising one or more links; receive a second one or more QoE metrics measured by a second probe traversing a second path comprising one or more links; determine, from the first one or more QoE metrics, that the first path has an anomaly; determine, from the second one or more QoE metrics, that the second path has an anomaly; and determine, in response to determining the first path and the second path has an anomaly, based on the type of metrics and the type of links, that an intersection between the first path and the second path is a root cause of the anomaly.

Abstract: An alarm service can receive an alarm rule as an “intent” that defines a rule in a high level “natural language.” An alarm rule compiler can receive the intent and translate the high level intent into one or more lower level rules that can be programmatically processed by multiple alarm rule execution engines. Devices in a network system can be associated with alarm rule execution engines in a distributed manner. For example, devices in a network can be associated with different instances of an alarm rule execution engine, thus distributing the resource usage for obtaining telemetry data and processing alarms with respect to the devices in a network across multiple alarm rule execution engines.

Abstract: This disclosure describes techniques that determine device connectivity in the absence of a network layer 2 discovery protocol such as Link Layer Discovery Protocol (LLDP). In one example, this disclosure describes a method that includes retrieving, from a bridge data store of a bridge device on a network having one or more host devices, a plurality of first interface indexes, wherein each first interface index corresponds to a network interface of network interfaces of the bridge device; retrieving, from the bridge data store, remote network addresses corresponding to the network interfaces of the bridge device, each remote network address of the remote network addresses corresponding to a second interface index; selecting a remote network address having a second interface index that matches the first interface index; determining a host device having the selected remote network address; and outputting an indication that the bridge device is coupled to the host device.

Abstract: This disclosure describes techniques for monitoring, scheduling, and performance management for computing environments, such as virtualization infrastructures deployed within data centers. In one example, a method includes obtaining, by a policy controller, a first profile for an element of a virtualization infrastructure, the first profile comprising a first ruleset having one or more alarms; obtaining, by the policy controller, a second profile for a group of one or more elements including the element, the second profile comprising a second ruleset having one or more alarms; modifying, by the policy controller based at least on the element being a member of the group, the first profile to generate a modified first profile comprising the first ruleset and the second ruleset; and outputting, by the policy controller to a computing device, the modified first profile.

Abstract: The disclosure describes techniques for network monitoring and fault localization. For example, a controller comprises one or more processors operably coupled to a memory configured to: receive a first one or more Quality of Experience (QoE) metrics measured by a first probe traversing a first path comprising one or more links; receive a second one or more QoE metrics measured by a second probe traversing a second path comprising one or more links; determine, from the first one or more QoE metrics, that the first path has an anomaly; determine, from the second one or more QoE metrics, that the second path has an anomaly; and determine, in response to determining the first path and the second path has an anomaly, based on the type of metrics and the type of links, that an intersection between the first path and the second path is a root cause of the anomaly.

Abstract: This disclosure describes techniques for presenting information about a network, virtualization infrastructure, cluster, or other computing environment, and may involve presentation of user interfaces that may enable nuanced, unique, and/or comprehensive insights into how infrastructure elements and computing resources are being used and information about patterns of usage and/or utilization. This disclosure also describes techniques for communicating, within a computing system, information used to create, update, and/or modify the user interfaces that present information about a network, virtualization infrastructure, cluster, or other computing environment.

Abstract: This disclosure describes techniques for monitoring, scheduling, and performance management for virtualization infrastructures within networks. In one example, a computing system includes a plurality of different cloud-based compute clusters (e.g., different cloud projects), each comprising a set of compute nodes. Policy agents execute on the compute nodes to monitor performance and usage metrics relating to resources of the compute nodes. Policy controllers within each cluster deploy policies to the policy agents and evaluate performance and usage metrics from the policy agents by application of one or more rulesets for infrastructure elements of the compute cluster. Each of the policy controllers outputs data to a multi-cluster dashboard software system indicative of a current health status for the infrastructure elements based on the evaluation of the performance and usage metrics for the cluster.

Abstract: The present invention addresses the need for improved virtualized cloud infrastructure policy implementation and management in order allow real-time monitoring and optimization of virtualized resources. It provides systems and methods for real-time cloud infrastructure policy implementation and management that include a plurality of host devices, a plurality of real-time probe agents associated with the plurality of host devices operating on each of the plurality of host devices, and a policy engine communicatively coupled to the plurality of host devices and containing a policy associated with an application program deployed in at least one of the plurality of host devices. The policy engine is programmed to monitor in real time changes in deployment of the application program across the plurality of host devices and to push the policy to the real-time probe agent operating on each host device on which the application program is deployed.

Abstract: Techniques to display a graphic representation of a computer network topology are described. In one example, a network device is configured to generate an output comprising a graphic representation of a topology of a computer network, the computer network comprising compute nodes interconnected by a packet-based communications network provided by a set of network devices, wherein the policy controller is further configured to: identify, amongst the compute nodes or the network devices, a network element having state information indicating an operational state of interest; modify state information for one or more resources that are coupled to the identified network element; and based upon whether the network topology information indicates a visible graphic symbol for the identified network element, output a visual indicator corresponding to the visible graphic symbol or corresponding to a visible resource that is communicably coupled to the identified network element.

Abstract: An alarm service can receive an alarm rule as an “intent” that defines a rule in a high level “natural language.” An alarm rule compiler can receive the intent and translate the high level intent into one or more lower level rules that can be programmatically processed by multiple alarm rule execution engines. Devices in a network system can be associated with alarm rule execution engines in a distributed manner. For example, devices in a network can be associated with different instances of an alarm rule execution engine, thus distributing the resource usage for obtaining telemetry data and processing alarms with respect to the devices in a network across multiple alarm rule execution engines.

Abstract: A telemetry service can receive telemetry collection requirements that are expressed as an “intent” that defines how telemetry is to be collected. A telemetry intent compiler can receive the telemetry intent and translate the high level intent into abstract telemetry configuration parameters that provide a generic description of desired telemetry data. The telemetry service can determine, from the telemetry intent, a set of devices from which to collect telemetry data. For each device, the telemetry service can determine capabilities of the device with respect to telemetry data collection. The capabilities may include a telemetry protocol supported by the device. The telemetry service can create a protocol specific device configuration based on the abstract telemetry configuration parameters and the telemetry protocol supported by the device. Devices in a network system that support a particular telemetry protocol can be allocated to instances of a telemetry collector that supports the telemetry protocol.

Abstract: This disclosure describes techniques for monitoring, scheduling, and performance management for virtualization infrastructures within networks. In one example, a computing system includes a plurality of different cloud-based compute clusters (e.g., different cloud projects), each comprising a set of compute nodes. Policy agents execute on the compute nodes to monitor performance and usage metrics relating to resources of the compute nodes. Policy controllers within each cluster deploy policies to the policy agents and evaluate performance and usage metrics from the policy agents by application of one or more rulesets for infrastructure elements of the compute cluster. Each of the policy controllers outputs data to a multi-cluster dashboard software system indicative of a current health status for the infrastructure elements based on the evaluation of the performance and usage metrics for the cluster.

Abstract: Techniques for resource monitoring and managed message reordering in a data center are described. In one example, a computing system comprises an ingress engine to receive a message from a network device in a data center comprising a plurality of network devices and the computing system; and in response to receiving the message from a network device in the data center, communicate the message to an appropriate collector application corresponding to the message's protocol type in compliance with at least one requirement for data stored in a message flow communicated from one or more network devices to the computing system.

Abstract: This disclosure describes techniques that include receiving underlay flow data from a network having an underlay network and one or more overlay network, storing information identifying, for each underlay data flow, an overlay network, displaying, within a display, a topological map of at least a portion of the underlay network, highlighting a data path through the displayed topological map, the highlighted data path extending through the underlay network from the underlay network source of the respective underlay data flow to the underlay network destination of the respective underlay data flow; receiving a request for metrics associated with the highlighted data path, wherein receiving a request includes receiving, via a graphical user interface, an indication selecting at least a portion of the highlighted data path; and displaying, proximate to the highlighted data path, metrics associated with data traffic through the selected portion of the highlighted data path.

Abstract: This disclosure describes techniques that determine device connectivity in the absence of a network layer 2 discovery protocol such as Link Layer Discovery Protocol (LLDP). In one example, this disclosure describes a method that includes retrieving, from a bridge data store of a bridge device on a network having one or more host devices, a plurality of first interface indexes, wherein each first interface index corresponds to a network interface of network interfaces of the bridge device; retrieving, from the bridge data store, remote network addresses corresponding to the network interfaces of the bridge device, each remote network address of the remote network addresses corresponding to a second interface index; selecting a remote network address having a second interface index that matches the first interface index; determining a host device having the selected remote network address; and outputting an indication that the bridge device is coupled to the host device.

Abstract: This disclosure describes techniques that include using an automatically trained machine learning system to generate a prediction. In one example, this disclosure describes a method comprising: based on a request for the prediction: training each respective machine learning (ML) model in a plurality of ML models to generate a respective training-phase prediction in a plurality of training-phase predictions; automatically determining a selected ML model in the plurality of ML models based on evaluation metrics for the plurality of ML; and applying the selected ML model to generate the prediction based on data collected from a network that includes a plurality of network devices.

Abstract: Techniques to display a graphic representation of a computer network topology are described. In one example, a network device is configured to generate an output comprising a graphic representation of a topology of a computer network, the computer network comprising compute nodes interconnected by a packet-based communications network provided by a set of network devices, wherein the policy controller is further configured to: identify, amongst the compute nodes or the network devices, a network element having state information indicating an operational state of interest; modify state information for one or more resources that are coupled to the identified network element; and based upon whether the network topology information indicates a visible graphic symbol for the identified network element, output a visual indicator corresponding to the visible graphic symbol or corresponding to a visible resource that is communicably coupled to the identified network element.