How to troubleshoot the Ops Agent

Yuri, identified in the transcript as an SRE at Google Cloud, decomposes missing VM telemetry into four boundaries: agent installation, service startup, source collection, and API delivery. This stage-by-stage model helps distinguish one failure class from another instead of treating absent logs or metrics as a single problem ([00:00:00]-[00:03:06]).

Key Points Covered

• Installation and startup are separate checks: In the demonstrated environment, unsupported operating systems, legacy-agent conflicts, or invalid configuration can stop the pipeline early ([00:01:01]-[00:02:04]).
• Collection modules have their own dependencies: Credentials, proxies, network paths, and API reachability can prevent logging or monitoring modules from starting ([00:02:04]).
• Inspect source and destination boundaries: The workflow checks module status, source readability, local metrics, module logs, API permissions, access scopes, and quotas ([00:03:06]).
• Missing data is not automatically data loss: The episode does not explain fleet-wide diagnosis or how to distinguish delayed ingestion from permanent loss.

This is a 2022 troubleshooting guide. Verify current supported systems, service and module names, commands, permissions, access-scope behavior, quotas, and legacy-agent compatibility; do not reconstruct exact errors from auto-captions.

Full video: https://www.youtube.com/watch?v=Sd0iznXSVcc(opens in a new tab)