Monitor it
Monitoring in 1.5 GiB of RAM
- Prometheus, Grafana, Loki and probes
- A Golden Signals dashboard
- Seven alerts, delivered by email
rttf · 3 recipes · monitoring 1.0.0 · S3 1.2.0 · backups 1.2.0
The three things a small k3s cluster needs after "it runs": knowing what it's doing, somewhere to put files and sites, and backups that have actually come back at least once. Three tested Flux recipes, one price.
What is in it
The S3 recipe creates the bucket the backups recipe needs, and the monitoring recipe adds dashboards and alerts for both. Each comes as its own zip with a README runbook, pinned versions and an AGENTS.md.
Made for AI coding agents
Each recipe ships with an AGENTS.md and a CLAUDE.md written as an operating procedure: what to ask you, what to check first, the exact steps, how to verify, and where to stop for your approval.
# following each recipe's AGENTS.md
✓ 1/3 S3 + static sites: buckets and keys created
✓ 2/3 backups: restore tests passed
✓ 3/3 monitoring: targets up, test alert emailed
? Each push to your Flux repo waits for your yes
Illustration of the steps in AGENTS.md, not a recorded session.
Tested, not promised
All three at once in a single cluster was not tested; it needs a bigger machine than the one used.
Questions
Yes. Each recipe is its own zip with a START-HERE.md, a README runbook, a VERSIONS.md and the agent files, so you can set them up one at a time.
S3 first, because it provides the bucket the backups recipe uses for Postgres. Then backups, then monitoring, which adds dashboards and alerts for the other two.
Partly. The S3 recipe can run on three nodes, with a copy on each, so any one node can go. Monitoring and backups run as single copies; an HA monitoring recipe is on the bench.
It runs on 64-bit ARM: it was tested from zero on an arm64 VM. It was not tested on a Pi itself, and 32-bit ARM is not supported.
Monitoring, storage and tested backups for a small k3s cluster, all in Git. Each recipe tested from zero on x86_64 and arm64.
Get all three €59