Changes to the process.command and process.command_line metrics for Agent-based Host Monitoring in SolarWinds Observability SaaS

Starting with uams-otel-collector-plugin v1.14.50 version and onwards, we are removing following attributes from process.*  metrics for Agent-based host monitoring:

• process.command
• process.command_line

Impact on the alerts:

Any alerts currently using the removed attributes in their conditions may behave differently after the update:

• For “process down” style conditions (for example, alerts that trigger when a process count drops to 0), the alert may fire once when the updated plugin is deployed and the time condition is met, because the underlying metric definition has changed.
• For “value higher than X” style conditions (for example, thresholds based on metrics filtered by these attributes), the alert may stop triggering entirely because the filter attribute is no longer present.

This change is required to fully comply with the latest OpenTelemetry Collector specification used to collect and process host metrics and to align with our ongoing security hardening efforts. 

You can use any of the following approaches, depending on your needs:

1. Temporarily pause auto-update for affected hosts scoped in the impacted alert (immediate mitigation if version of uams-otel-collector-plugin used is below v1.14.50). If you need more time to adjust your alerts, you can disable auto-update for the impacted hosts’ UAMS agent/plugin so they remain on the current plugin version while you update alert configurations. Once your alerts are updated, you can re-enable auto-update. Auto-update feature can be enabled/disabled in the UI via Settings >> Agents >> Auto Update Column >> toggle button.
2. Update alert conditions to use supported attributes:
   1. Use process.executable.name  as the primary attribute for identifying processes in alert conditions going forward.
   2. For long-running processes where you need to distinguish specific instances, you can additionally use process.pid in your conditions.
3. For Windows hosts where you want more precise status monitoring of services, you can use the Services sub‑tab under the Processes tab in the entity details view and create an alert directly.  
4. If you have a strict requirement to keep using process.command_line or similar detail, you can instead deploy the SolarWinds OTel Collector with a custom host integration configuration that continues to collect command-line data:
   1. SolarWinds OTel Collector overview and deployment options
   2. Example host monitoring configuration for the SolarWinds OTel Collector