Smarts can be configured to use one of the operator, MSP, or enterprise user for discovering and monitoring the VeloCloud orchestrator. Also, there is a flexibility to opt for only a subset of Tenants when using the MSP and operator user. When a subset of interested tenants is specified using a regex, the discovery and monitoring only deal with VeloCloud topology that belongs to these Tenants.

In summary Smarts provides:

  • Support for operator, MSP, and enterprise user types.
  • Enterprise filtering for operator and MSP users.

Table Parameters and Descriptions illustrates about the new parameters and their descriptions in the ESM file.

Table 1. Parameters and Descriptions
Parameters Descriptions
VCO_USERTYPE-<Orchestrator_host_IP_address> operator|enterprise|MSP Defines the VeloCloud user type.
Note: User type is a mandatory parameter and it can be:
  • operator

    or

  • enterprise

    or

  • MSP
VCO_ENTERPRISE_FILTER-<Orchestrator_host_IP_address> <REGEX FILTER>

Defines the RegEx pattern for list of enterprises in the VeloCloud that ESM discovers and monitors.

The regex pattern must confirm to patterns defined in java.util.regex documentation. This is an optional parameter.

Default filter is .* (i.e discover all enterprises that belong to a specific user).

For example, if ESM discovers two Tenants, "A" and "B", then the parameter can be defined as "A|B".

VCO_ENTERPRISE_FILTER-<Orchestrator_host_IP_address> A|B

VCO_HTTP_RESPONSE_TIMEOUT-<Orchestrator_host_IP_address> <Time in Seconds>

Defines the http timeout for response in seconds for VeloCloud discovery and monitoring collector while fetching Topology using REST API.

This is an optional parameter. Minimum, Default, and Maximum values are; 120, 240 and 600 respectively.

VCO_DISCOVERY_TIMEOUT-<Orchestrator_host_IP_address> <Time in Seconds>

Defines the discovery collector timeout in seconds for VeloCloud discovery.

This is an optional parameter. Minimum, Default, and Maximum values are; 3600, 15000 and 36000 respectively.

VCO_COLLECTOR_DISCOVERY_LOG_LEVEL-<Orchestrator_host_IP_address> <LOGLEVEL>

Defines the VeloCloud collector logging level for discovery.

Valid debug levels are:

WARNING,SEVERE,OFF,INFO,FINEST,FINER,FINE,CONFIG,ALL

This is an optional parameter and if not defined, the default loglevel value is INFO.

VCO_COLLECTOR_MONITORING_LOG_LEVEL-<Orchestrator_host_IP_address> <LOGLEVEL>

Defines the VeloCloud collector logging level for monitoring.

Valid debug levels are:

WARNING,SEVERE,OFF,INFO,FINEST,FINER,FINE,CONFIG,ALL

This is an optional parameter and if not defined, the default loglevel value is INFO.

VCO_COLLECTOR_MAX_HEAP_MEMORY-<Orchestrator_host_IP_address> <MaxHeapSize>

Defines the VeloCloud collectors maximum heap memory size (in Mega Bytes) for discovery and monitoring.

This is an optional parameter. Minimum, Default, and Maximum values are; 64, 2048, and 4096 respectively.

#VCO_KAFKA_ADAPTER_QUEUE_HIGH_WATERMARK <integer value>

By default, ESM reads all the available messages in the kafka bus. The read messages (discovery/monitoring) are stored in ESM process memory.

If the processing of these message by ESM is slower than the message producer then it can result in higher memory foot print for ESM.

This default behavior can be changed by setting a high and low water mark for the kafka adapter internal queue. Both high and low must be set in this case.

VCO_KAFKA_ADAPTER_QUEUE_HIGH_WATERMARK parameter is for specifying the maximum size of the kafka adapter internal queue size at which the consumer is stopped to minimize the memory footprint of the ESM server.

For example, VCO_KAFKA_ADAPTER_QUEUE_HIGH_WATERMARK 50000 can ensure that after beyond 50k messages being in the kafka adapter’s internal buffer the consumption of further messages from the kafka bus is stopped.

VCO_KAFKA_ADAPTER_QUEUE_LOW_WATERMARK <integer value>

Specifies the minimum size of the kafka adapter internal queue size at which the consumer is restarted.

Refer the description of the VCO_KAFKA_ADAPTER_QUEUE_HIGH_WATERMARK. Both high and low water mark must be specified. By default, this is disabled and in this case the kafka adapter makes reads all the available messages from the kafka bus.

For example, VCO_KAFKA_ADAPTER_QUEUE_LOW_WATERMARK 10000 restarts the message consumption from the kafka bus once the internal queue size drops below 10k.

Note: The following parameters must be modified in esm-param.conf file to discover a Topology having around 3500 Vedges, the parameters must be modified in esm-param-conf file based on user deployment:
  • MessagePollTimeoutPeriodInSeconds-<kafka_IP_Address> 2400

    When the latency between orchestrator and the DCF collector is higher, set the flag to a higher value. This flag dictate the timeout value for the ESM to wait for messages from kafka. If the timeout expires, then the discovery stops abruptly.

    Minimum, Default, and Maximum values are; 60, 1200 and 3000 respectively.

  • VCO_HTTP_RESPONSE_TIMEOUT-<Orchestrator IP> 240

    The timeout value for the http request that the DCF collector sends to the orchestrator. While configuring this value, the processing time of the orchestrator must also be considered into account as some of the REST API takes longer time to return.

  • VCO_DISCOVERY_TIMEOUT-<Orchestrator IP> 18000

    Increase the value of this parameter based on the topology size. This is the maximum time in seconds that the collector runs before stopping abruptly. This is a safeguard measure to prevent the collector from running under a false condition.