Monitoring and Health

Current infra-ansible monitoring baseline is metrics-first.

Typical stack

• Grafana Alloy for scrape + remote_write
• node_exporter for host metrics
• cAdvisor for container metrics
• SynckHub/Common app metrics endpoints

Design constraints

• metrics endpoints stay loopback-only
• outbound push model (no public inbound scrape)
• stable relabeling for instance and job

Minimum alert set

Start with absence and infrastructure pressure alerts:

• missing data for SynckHub/Common jobs
• disk usage thresholds
• host CPU and memory pressure

Health objective

You should be able to answer quickly:

• Is service up?
• Is it overloaded?
• Is authz/provision sync stale?