Onboarding NPIs to Metric Collector Module

This document describes how to enable and configure the tlBMC Metric Collector Module for an NPI. This module collects software-defined metrics such as socket statistics and netfilter connection statistics.

Prerequisites

The tlBMC Metric Collector Module must be enabled for your platform in tlbmc_config_bundle.textproto. If it is not enabled, tlBMC will not collect metrics. See tlbmc_config_bundle.textproto in tlbmc source.

To enable the module, add metric_collector_module { enabled: true } to your platform's configuration in platform_to_config map. For example:

platform_to_config {
  key: "your_platform_name"
  value {
    metric_collector_module {
      enabled: true
    }
  }
}

Metric Configuration

The Metric Collector Module requires a configuration file named software_metrics_config.textproto to specify which metrics to collect. This file should be placed in your platform's configuration directory, typically here.

The software_metrics_config.textproto file should contain a SoftwareMetricsConfig message, as defined in software_metrics_config.proto.

The SoftwareMetricsConfig message allows you to specify ports for socket statistics and netfilter statistics:

• socket_stat_ports: A list of TCP ports for which to collect socket statistics (e.g., number of connections in various TCP states).
• netfilter_ports: A list of TCP ports for which to collect netfilter connection tracking statistics.

Example Configuration

Here is an example of software_metrics_config.textproto that configures collection for SSH (port 22) and HTTPS (port 443):

# proto-file: third_party/milotic/external/cc/tlbmc/configs/software_metrics_config.proto
# proto-message: SoftwareMetricsConfig

socket_stat_ports: 22
socket_stat_ports: 443
netfilter_ports: 22
netfilter_ports: 443

Collected Metrics

When enabled and configured, the module collects metrics at a default interval of 30 seconds. The collected metrics can be accessed via tlBMC's Redfish interface at /redfish/v1/Managers/bmc/ManagerDiagnosticData/Oem/Google/BmcSocketMetrics/, and /redfish/v1/Managers/bmc/ManagerDiagnosticData/Oem/Google/BmcNfMetrics/.

Example Output

Example output of /redfish/v1/Managers/bmc/ManagerDiagnosticData/Oem/Google/BmcSocketMetrics/

{
  "Port": {
    "443": {
      "CLOSED": 0,
      "CLOSE_WAIT": 0,
      "CLOSING": 0,
      "ESTABLISHED": 1,
      "FIN_WAIT_ONE": 0,
      "FIN_WAIT_TWO": 0,
      "LAST_ACK": 0,
      "LISTENING": 0,
      "SYN_RECEIVED": 0,
      "SYN_SENT": 0,
      "TIME_WAIT": 0
    }
  },
  "Timestamp": 1700000000
}

Example output of /redfish/v1/Managers/bmc/ManagerDiagnosticData/Oem/Google/BmcNfMetrics/

{
  "Port": {
    "443": {
      "Connections": 12345
    }
  },
  "Timestamp": 1700000000
}