Your fleet.
eyes everywhere.
Deploy across all your machines. ilArgus shows you what is healthy, what is failing, and what is about to. It gives you and your AI agents instantaneous access to every machine, with all the controls.
ilargus · installSTANDBY
$
resolving ilargus.io … ok
fetching ilargus 0.1.1 (linux/amd64) … 8.2 MB
installing → /usr/local/bin/ilargus
✓ eye ready ilargus eye
✓ head ready ilargus head
ilArgus installed. run: ilargus head
ilargus headdeploying eyes →
dev-eyemacOS 14 · arm64
connecting…
web-01ubuntu 22.04 · amd64
connecting…
db-02debian 12 · amd64
connecting…
edge-03alpine 3.19 · arm64
connecting…
0 eyes online· streaming to head
ARGUS
Fleet
▸ argus / fleet / ● dev-eye
dev-eye
ONLINEMacBookPro · last seen 1s
CPU
0.0%
Memory
0.0%
12.1 / 16.0 GiB
TerminalStatsContainers · 0Inspect
CPU
0.0%
Memory
75.7%
Swap
79.2%
Load avg5.06 5.37 4.65
Network↓ 42.0 ↑ 41.3 KiB/s
live logs · dev-eye
14:22:01INFOeyeheartbeat ok · cpu 23.1% · mem 75.7%
14:22:01INFOdockercontainer web healthy · up 8h
14:22:02OKheadfleet synced · 4 eyes online
14:22:02WARNeyemem 75.7% approaching soft limit
14:22:03INFOdb-02pg checkpoint · 1.2k tx/s
14:22:03INFOeyenet ↓ 42.0 ↑ 41.3 KiB/s
dev-eye $
CONTAINER IDIMAGESTATUSNAMES
e3f1a9c2b4d8nginx:1.27Up 8hweb
7a2c5e1f9b03postgres:16Up 8hdb
b91d4f6a2c70redis:7-alpineUp 3hcache
One binary · three modes
head
ilargus headCentral control. Receives all data from your Eyes and gives you full control over all of them.eye
ilargus eyeRuns on each machine. Connects to the Head, sends back metrics and logs, and lets you control it.cli
ilargusYour terminal. ssh, logs, exec, vnc, deploy.01 / see
See everything.
Metrics, logs, terminals, containers, remote desktops, Kubernetes. ilArgus draws your entire fleet onto a single surface. No juggling a dozen tools, no stitching tabs together. The whole picture, in one place, the moment you open it.
ilargus head · prod / fleet / ilargus-eye-07LIVE · 0.4 s
cpu · 60 s · per-eye p95streaming
eyes · top by load
hoststatecpumemage
prod-edge-01online71%88%0s
prod-edge-04online62%74%0s
do-droplet-03degraded——12s
ilargus-eye-07online32%61%0s
ilargus-eye-12online18%44%1s
mac-build-02offline——4m
logs · all eyes12.4k / s
14:32:11eye-07infocontainer nginx-prod started (pid 184)
14:32:11edge-01okhealth check passed (12 ms)
14:32:10droplet-03warnvolume mount /data degraded — readonly
14:32:10eye-12infopostgres-13 vacuum complete (4.2 MiB freed)
14:32:09edge-04errrequest id e3b0c… 502 upstream timeout
14:32:09eye-07infometrics flushed · 412 samples · 8.2 ms
14:32:08droplet-03errcontainer api-v2 exited with code 137
14:32:08edge-01okreload SIGHUP accepted · workers 8 → 8
14:32:07eye-12infotap container redis-cache stdout · 14 lines
containers · ilargus-eye-07
nginx-prod · ingress0.4 %38 MiB
api-gateway · ingress8.2 %214 MiB
postgres-13 · data3.1 %1.2 GiB
redis-cache · data0.6 %82 MiB
worker-jobs · workers14 %412 MiB
api-v2 · ingress— exit 137
shell · ilargus-eye-07 · ssh broker · 13 ms rttconnected
root@ilargus-eye-07 ~ # ilargus tail --container=api-v2 --since=1m
[14:32:08] api-v2 volume mount /data is readonly: cannot write to /data/queue.lock
[14:32:08] api-v2 graceful shutdown requested · SIGTERM
[14:32:08] api-v2 → exited code 137 after 0.62 s
root@ilargus-eye-07 ~ # ilargus exec api-v2 -- df -h /data
02 / act
Act instantly.
Zero-setup SSH into any machine. Remote desktop access over VNC. Docker actions and Kubernetes control at your fingertips. The distance between noticing a problem and fixing it collapses to a single click.
your laptop · ~/workRTT 13 ms
$ ilargus ssh prod-edge-04
↳ resolving via head · broker selected · session opened (13 ms)
✓ connected as root@prod-edge-04 · 11d 06h uptime
root@prod-edge-04 # ilargus restart nginx-prod
✓ nginx-prod · stopped · started · health passed (44 ms)
root@prod-edge-04 # ilargus exec api-v2 -- sh
↳ docker exec -it 8a2c1f sh · attached
/app # ^D
root@prod-edge-04 # ilargus vnc mac-build-02 --port=5900
✓ tunnel up · vnc://localhost:5900 (opening)
⌘Krestart nginxESC TO CLOSE
restart container ·
nginx-prod · prod-edge-04↵ runrestart container ·
nginx-prod · prod-edge-01↵restart pod ·
nginx-ingress-* · k8s/prod↵view config ·
/etc/nginx/nginx.conf↵↑↓navigate↵run⌃Rretry on all
Open a shell
ilargus ssh <host> · zero config, broker-routed
Open remote desktop
ilargus vnc <host> · tunneled, encrypted
Run on every host
ilargus run --all -- <cmd> · parallel, with bail-out
03 / delegate
Used by your AI agents.
Plug in the ilArgus skill or the ilArgus MCP and your agents take the wheel. Your fleet, managed in plain language.
user
>Claude, my DigitalOcean droplet is down, find and fix it.
claude
Looking at your fleet.
ilargus
eyes --state=offline,degraded1 result · do-droplet-03ilargus
ssh do-droplet-03 -- ilargus ps --state=exitedapi-v2 · exit 137ilargus
logs do-droplet-03/api-v2 --since=2mvolume /data readonly · OOM cgroupilargus
exec do-droplet-03 -- mount -o remount,rw /dataok · 22 msilargus
restart do-droplet-03/api-v2started · health passed (44 ms)A container crashed on a misconfigured volume. Fixed, restarted, production is back up.
do-droplet-03 · online · 12 s ago
ilArgus skill
drop-in for Claude · 14 verbs
ilArgus MCP
Model Context Protocol · streaming, structured
04 / open
Fully open. Built in Rust.
One 8 MB binary. 45 MB of RAM. A thousand machines streaming metrics and logs without breaking a sweat. Fast, light, and fully open source, with nothing hidden and nothing held back.
single binary8mbStatic. Stripped. No runtime. Same binary runs as head, eye, or cli — selected by the first argument.
per-eye footprint45mb ramResident. Idle. Including its built-in metrics scraper, log tap, and shell broker.
per-head capacity1k+ hostsOn a single small VM. Metrics and logs streaming in parallel. Tested, not promised.
Apache-2.0github.com/ilargus/ilarguswritten in Rust 2024no telemetry · no phone-homeaudited build · reproducible
Deploy in one command.
Watch your fleet come online.
$curl -sSf https://ilargus.io/install | sh