A small set of systems I’ve built or reshaped. Public where possible. Private when required.
Private
CRM / Automation
Event-driven workflow automation tied to real CRM state
What it is: lightweight services that react to CRM events and keep downstream work in sync without manual chasing.
How it works:
- Triggers from explicit CRM events (not schedule-based sweeps)
- Normalizes inputs and writes back clear, auditable state
- Sends a small number of owned alerts tied to failure modes
Why it exists:
- Reduced missed handoffs and “someone usually checks that” work
- Improved consistency by making ownership and state explicit
Demo (private)
Private
State / Sync
Vendor status normalization and verification pipeline
What it is: a source-of-truth status model that replaces email-thread state with consistent transitions.
How it works:
- Ingests updates from vendors and internal events
- Normalizes to a small set of states with clear meanings
- Syncs updates back into the CRM automatically
Why it exists:
- Fewer one-off checks and “where is this at?” conversations
- Status became reliable enough to automate downstream steps
Screenshots (redacted)
Private
Alerting
Condition-based alerts that stay actionable
What it is: alerting designed around specific failure modes, not broad reminders.
How it works:
- Triggers only when a defined state + required fields are met
- Messages include the minimum context to take action
- Deduping and audit logging prevent spam
Why it exists:
- Reduced alert fatigue and increased follow-through
- Made “exceptions” visible instead of hidden in DMs
Demo (private)
Private
Reliability
Reliable signals for small services
What it is: a minimal set of checks around real failure points: health, latency, and stale state.
How it works:
- Simple health endpoints per service
- Latency monitoring for key dependencies
- “Stale state” detection for workflows that silently drift
Why it exists:
- Faster detection without dashboard overload
- More confident changes because regressions surfaced quickly
Screenshots (redacted)
Public
Infra / Static
ericminish.com bring-up
What it is: a static site served from Proxmox behind Cloudflare Tunnel with health checks and simple rollback.
How it works:
- Static files served by a minimal web stack
- Health endpoint for basic availability checks
- Ops baseline: onboot, backups, snapshots
Why it exists:
- Stable foundation for iteration without platform churn
Repo (public)
Public
Homelab
Homelab: constrained reliability practice
What it is: a small environment to validate monitoring, upgrades, and recovery patterns safely.
How it works:
- Proxmox nodes hosting a few purpose-driven services
- DNS filtering and latency monitoring as core signals
- Iterates slowly to keep it maintainable
Why it exists:
- Turns “best practices” into muscle memory
Details
