Health-based management of a network
In some examples, a system stores, in a first database having a first schema, metrics received from a network comprising communication nodes, the metrics relating to operations of the communication nodes, and the first database associating the metrics with metadata corresponding to hierarchical components in a topology of the network. In response to an alert relating to an issue in the network, the system computes a health measure based on the metrics, the health measure indicating a health status of a first component, performs a dynamic runtime mapping of the metadata associated with the metrics in the first database having the first schema with corresponding metadata in a second database having a second schema different from the first schema, where the second database contains information of a topology of the network, and initiates a management action to address the health status based on the health measure and the dynamic runtime mapping.
1 . A non-transitory machine-readable storage medium comprising instructions that upon execution cause a system to:
store, in a first database having a first schema, metrics received from a network comprising communication nodes, the metrics relating to operations of the communication nodes, and the first database associating the metrics with one or more tags corresponding to hierarchical components in a topology of the network;
store, in a second database having a second schema different than the first schema, an information of the topology of the network, comprising an indication of the hierarchical components and how the hierarchical components are related to one another;
receive an alert relating to an issue in the network, the alert comprising an identifier of a first component of the hierarchical components that is associated with the alert; and
in response to the alert:
trigger collection of additional metric data at a first sampling rate, wherein the metrics are received from the network at a second sampling rate less than the first sampling rate;
compute a health measure based on the metrics and the additional metric data, the health measure indicating a health status of the first component,
perform a dynamic runtime mapping between the first schema and the second schema, based upon schema mapping information indicating one or more keys of the second schema that correspond to the one or more tags of the first schema;
correlate one or more tag values of the one or more tags in the first database that are associated with a subset of the metrics corresponding to the identifier of the first component with corresponding one or more key values in the second database that identify the first component in to associate the subset of the metrics with a corresponding subset of the information of the topology of the network contained by the second database;
identify a subset of the hierarchical components affected by the alert from the information of the topology of the network stored in the second database, based upon the corresponding subset of the information of the topology of the network;
initiate a management action on one or more of the subset of the hierarchical components affected by the alert to address the health status based on the health measure and the corresponding subset of the information of the topology of the network;
identify that the issue in the network has been resolved; and
in response to identifying that the issue in the network has been resolved, decrease the first sampling rate.
2 . The non-transitory machine-readable storage medium of claim 1 , wherein the alert is based on local processing of metrics by one or more of the communication nodes.
3 . The non-transitory machine-readable storage medium of claim 1 , wherein the additional metric data comprises a different metric type than a metric type of the metrics received from the network prior to the alert and the instructions upon execution cause the system to:
create, at runtime, a schema to store the additional metric data that comprises the different metric type.
4 . The non-transitory machine-readable storage medium of claim 1 , wherein the identified subset of the hierarchical components affected by the alert comprises at least one hierarchical component other than the first component.
5 . The non-transitory machine-readable storage medium of claim 1 , wherein the instructions upon execution cause the system to:
access information associating a list of metrics with the first component; and
identify one or more metrics to collect based on the list of metrics,
wherein the triggering of the collection of the additional metric data is based on the identified one or more metrics.
6 . The non-transitory machine-readable storage medium of claim 1 , wherein the first database having the first schema comprises a timeseries database.
7 . The non-transitory machine-readable storage medium of claim 6 , wherein the second database having the second schema is selected from among a relational database having a relational schema, a graph having a graph-based schema, or a database having a document-based schema.
8 . The non-transitory machine-readable storage medium of claim 1 , wherein the instructions upon execution cause the system to:
initiate at least one of: adjusting forwarding of data units by a switch or group of switches or setting a port to an inactive state as the management action.
9 . The non-transitory machine-readable storage medium of claim 1 , wherein the hierarchical components comprise at least one of: a port of a communication node, a communication node, or a group of communication nodes.
10 . A method comprising:
receiving, at a system comprising a hardware processor, metrics and events from communication nodes in a network, wherein the events are generated by the communication nodes based on local processing of the metrics at the communication nodes;
storing, by the system, the metrics and the events in a first database having a first schema;
storing, in a second database having a second schema different than the first schema, an information of a topology of the network, comprising an indication of hierarchical components of the network and how the hierarchical components are related to one another;
in response to an event comprising an alert relating to an issue in a component of the hierarchical components of the network, the alert comprising an identifier of the component:
triggering, by the system, one or more communication nodes in the network to collect additional metric data associated with the component of the hierarchical components of the network at a first sampling rate, wherein the metrics are received from the communications nodes in the network at a second sampling rate less than the first sampling rate;
performing, by the system, dynamic runtime mapping between the first schema and the second schema, based upon schema mapping information indicating one or more keys of the second schema that correspond to one or more tags of the first schema;
correlating one or more tag values of the one or more tags of the first schema associated with a subset of the metrics and the events, in the first database, corresponding to the identifier of the component of the hierarchical components of the network with corresponding one or more key values in the second database that identify the component of the hierarchical components of the network to associate the subset of the metrics and the events with a corresponding subset of the information of the topology of the network contained by the second database;
identifying a subset of the hierarchical components affected by the alert from the information of the topology of the network stored in the second database, based upon the corresponding subset of the information of the topology of the network;
initiating, by the system, a management action on the subset of the hierarchical components affected by the alert to address the issue based on the corresponding subset of the information of the topology of the network;
identifying that the issue in the network has been resolved; and
in response to identifying that the issue in the network has been resolved, decreasing the first sampling rate.
11 . The method of claim 10 , wherein the first schema is a timeseries database, and the second schema is selected from among a relational schema, a graph-based schema, or a document-based schema.
12 . A system comprising:
a processor; and
a non-transitory storage medium storing instructions executable on the processor to:
store metrics and events received from communication nodes in a network in a first database having a timeseries schema;
store, in a second database having a relational schema different than the timeseries schema, an information of a topology of the network, comprising an indication of hierarchical components of the network and how the hierarchical components are related to one another;
in response to an event comprising an alert relating to an issue in a component of the hierarchical components of the network, the alert comprising an identifier of a component of the hierarchical components that is associated with the alert:
trigger one or more communication nodes in the network to collect additional metric data corresponding to the component of the hierarchical components of the network at a first sampling rate, wherein the metrics are received from the communications nodes in the network at a second sampling rate less than the first sampling rate;
perform dynamic runtime mapping between the timeseries schema and the relational schema, based upon schema mapping information indicating one or more keys of the relational schema that correspond to one or more tags of the timeseries schema;
correlate one or more tag values of the one or more tags of the timeseries schema associated with a subset of the metrics and the events, in the first database, corresponding to the identifier of the component of the hierarchical components of the network with corresponding one or more key values in the second database that identify the component of the hierarchical components of the network to associate the subset of the metrics and the events with a corresponding subset of the information of the topology of the network contained by the second database;
identify a subset of the hierarchical components affected by the alert from the information of the topology of the network stored in the second database, based upon the corresponding subset of the information of the topology of the network;
initiate a management action on the subset of the hierarchical components affected by the alert to address the issue based on the corresponding subset of the information of the topology of the network;
identify that the issue in the network has been resolved; and
in response to identifying that the issue in the network has been resolved, decrease the first sampling rate.
13 . The system of claim 12 , wherein the tags and the keys comprise identifiers of components in the network.
14 . The non-transitory machine-readable storage medium of claim 1 , wherein the instructions upon execution cause the system to:
initiate the management action by instructing a control plane of the network to perform a control plane action comprising at least one of: updating forwarding information in one or more switches of the network, adjusting an activation status of a port or lane of links in the network, or adjusting a configuration of the network.