Sentinel Dead-Man / Coverage Health
Pilot Designed to notice when a Sentinel stops reporting, so silence is not mistaken for health.
- Watches
- Sentinel check-ins, expected coverage, stale data
- Designed to detect
- A Sentinel that has gone quiet, fallen behind, or left an expected area without coverage.
- Evidence
- Expected check-ins compared with the check-ins actually received.
- Vigil behavior
- Observe and report
Backup Runtime & Success Sentinel
Planned Designed to notice a backup process becoming less healthy before the day it fails completely.
- Watches
- job outcome, duration, retries, missed windows
- Designed to detect
- Backups becoming slower, less reliable, or missing their window.
- Evidence
- Backup job history, duration, and outcome over time.
- Vigil behavior
- Recommend
Storage Capacity & Latency Trend Sentinel
Planned Designed to spot storage that is filling or slowing at a rate that points to a future problem.
- Watches
- free capacity, growth rate, storage latency
- Designed to detect
- Capacity or latency trends heading toward a limit.
- Evidence
- Capacity and latency measurements trended over time.
- Vigil behavior
- Recommend
Service Flapping Sentinel
Planned Designed to notice a service that keeps restarting, even when it looks up at any given moment.
- Watches
- service restarts, state changes, uptime
- Designed to detect
- Repeated restarts or state changes over hours or days.
- Evidence
- Service state history.
- Vigil behavior
- Bounded action, where authorized
Certificate / Credential Expiry Sentinel
Planned Designed to give early notice of certificates and credentials approaching expiry.
- Watches
- certificate expiry dates, credential expiry dates
- Designed to detect
- Certificates or credentials nearing expiry.
- Evidence
- Expiry information read from certificates and credential metadata.
- Vigil behavior
- Bounded action, where authorized
Host Health Sentinel
Planned Designed to notice resource pressure or hardware warnings developing on a host.
- Watches
- CPU, memory, disk, hardware status
- Designed to detect
- Sustained resource pressure or emerging hardware warnings.
- Evidence
- Host resource and hardware indicators over time.
- Vigil behavior
- Recommend
Virtualization Health Sentinel
Planned Designed to check that virtual machines and their hosts are in their expected state.
- Watches
- virtual machine state, host state, resource contention
- Designed to detect
- Virtual machines not in their expected state, or contention building across hosts.
- Evidence
- Virtualization state compared with the expected configuration.
- Vigil behavior
- Recommend
UPS / Power Health Sentinel
Planned Designed to detect battery or power degradation before it affects continuity.
- Watches
- battery health, estimated runtime, on-battery events
- Designed to detect
- Declining battery health or reduced runtime.
- Evidence
- UPS status readings and power events.
- Vigil behavior
- Recommend