The sentinel for Proxmox: Atlas Watch
Monitoring on Proxmox usually turns into a project of its own: install Grafana, feed InfluxDB, hand-build the dashboards. The result looks good, but maintaining it becomes another job. Atlas Watch takes a different path: monitoring comes built into the cockpit already in use. Checks run on their own, and word arrives when something goes wrong, with no separate stack to set up.
Why a separate monitoring stack is a burden
The Grafana + InfluxDB + exporter chain is powerful, but setting it up, updating it and maintaining the dashboards is a load of its own. In a homelab most people set it up once, then it goes stale.
When monitoring lives in a separate tab, seeing a problem and fixing it happen in two different tools. The alert is in one place, the fix in another.
The worst is a silent failure: a backup fails overnight, a pool fills, a service enters a restart loop. Because no one is watching, it surfaces not in daylight but once it is too late.
What Atlas Watch does
It keeps a finger on the host’s pulse and surfaces what matters:
Scheduled health checks
A ready catalog of checks runs at set intervals (disk health, pool state, services, capacity and more) with no setup at all.
Instant email on critical findings
When a check turns critical, an email arrives without delay, with a smart cooldown so repeats do not pile up.
Notification center
Every finding gathers in one place; each alert shows where it is on a real mini-topology, not an abstract line.
Daily digest
Non-critical items accumulate into a digest: no email for every little thing, the day’s picture in one pass.
Flexible delivery
Email can be sent locally direct, or go out through a separate SMTP server; the best path is chosen for the setup at hand.
Custom rules
Beyond the ready catalog, custom thresholds and rules can be defined; Watch watches whatever it is told to watch.
Frequently asked questions
- Do I need to install Grafana or InfluxDB for Atlas Watch?
- No. Watch comes built into Atlas; a separate monitoring stack is not needed. Checks and alerts work right out of the box.
- How do alerts reach me?
- Critical findings arrive by email immediately; non-critical ones accumulate as a daily digest. Email can be sent locally or go out through a separate SMTP server. Notifications also gather in the notification center in the panel.
- Does it work without internet?
- Yes. Watch runs locally on the host; as long as it can send email locally or through a separate SMTP server, it needs no internet.
- Will the same alert bombard me all day?
- No. Critical alerts have a smart cooldown; the same finding does not produce the same email over and over. Repeats still show, but the inbox does not fill up.
- Can I choose what it watches?
- Yes. Alongside the ready check catalog, custom thresholds and rules can be defined; which finding is critical and which is a digest item is a setting.
- Does Watch replace a monitoring tool like Pulse?
- Different jobs. Monitoring tools specialize in dashboards; Watch is the sentinel of the cockpit already in use, sitting next to updates, storage and access in the same product. If monitoring is the main job, a dedicated tool makes sense; for managing the host from one place, Watch is already there.
Related entries
- The backup job stopped quietly: the costliest failure is not the loud one A backup that fails loudly gets fixed the same day. A backup that stops quietly is discovered on the day it was needed. The difference is not in a notification, it is in what you look at.
- What should the first screen of a Proxmox server tell you? Most overview screens answer "what exists". The useful question is "is anything wrong, and does it need me today". Those are different screens, and only one of them is worth opening every morning.
- Do you need a metrics stack, or does the built-in view already answer your question? Alerts and metrics are different tools for different questions. The built-in view answers "is something wrong now" perfectly well. A metrics stack answers "what happened last Tuesday at three", and that is the only reason to run one.
- Something happened last night: where the record actually is Alerts tell you that something happened. Logs tell you why. The catch is that the record you most need covers the moment the machine died, and on a default install that is the one most likely to be missing.