The platform main line survives a boot, and 8200 is the team line
Luna's go-live (2026-09-28) wrote [kamo-main-luna] into extensions_custom.conf and KAMO_LUNA_EXT / KAMO_LUNA_MAX_RING into globals_custom.conf by hand on the ho...
An AI workstation sleeps with its desktop stopped, and wakes to a fresh one
Measured on a throwaway -ai4 workstation (SP98 Task 16), the wake stalled because Xorg does not survive a resume: with no KMS driver it drives the console throu...
A quiet meeting keeps its AI (media_timeout 20m), again
Reapplies 960331f, reverted in 5d70873 on the wrong suspicion: the RTP loss seen right after that deploy was fail2ban's ban on 47.181.8.84 (fixed in the previou...
Livekit-sip's RTP never takes the router, and its hairpin is never banned
FreePBX names its public address (pbx.k1.kluster.kamocrm.com) in the SDP and the Contact it gives livekit-sip, which is outside local_net, so RTP and in-dialog ...
The heap fits the container, so a meeting is never OOMKilled
The image starts java -Xmx${JIGASI_MAX_MEMORY:-3072m} and the container was limited to 1Gi: the heap could outgrow the container, and Jigasi was OOMKilled mid-m...
Permanent forwards expire from the browser cache after an hour instead of never
From 2026-09-23 to 09-27 the kamouniverse-to-kamocrm redirectRegex 301'd kamouniverse.com to www.kamocrm.com. Traefik sends such a redirect with no Cache-Contro...
The meeting bridge trusts LiveKit from k3m1's node address too
livekit-sip is hostNetwork on k3m1, and k3m1 reaches 192.168.4.22 over its eno50 cable (192.168.4.22 via 10.0.50.0 dev eno50 src 10.0.50.1), so FreePBX sees eve...
An AI workstation goes to sleep with nothing on its desktop drawing
A page still drawing when the sleep image is written made the wake take minutes. SP11 Task 30's workstation slept with Kamo's sign-in page open (a full-screen z...
Never ban k3m1 — its node network joins the fail2ban allowlist
FreePBX runs with hostNetwork on k1m1. A pod on k3m1 reaches it SNATed to k3m1's InternalIP, 10.0.50.1, which none of the [DEFAULT] ignoreip ranges covered (the...
A lone call waits five minutes for company, not thirty seconds
Jigasi hangs up a call that nobody else joins within JIGASI_JVB_TIMEOUT (the image's default 30 s): Jicofo invites a participant only once there are two. An AI ...
Docs trusts api.kamouniverse.com as a WOPI host
A member signed in on internal.kamouniverse.com (the platform's primary apex since 2026-09-28) opens documents with **************** DocsService takes the last ...
Kamouniverse.com refuses unknown recipients at RCPT instead of bouncing them
kamouniverse.com, the platform's primary apex, is an alias domain of kamocrm.com: every <x>@kamouniverse.com is the <x>@kamocrm.com mailbox. It was mapped by a ...
The desktop-ws route reaches only the terminal socket, not every SecurityService path
**************** matched PathPrefix(`/desktop-ws`) on internal.* and stripped the prefix, so **************** reached ANY SecurityService path from the internet...
A caller an AI member transfers rings the person from its own channel
FreePBX sends every blind transfer to its global TRANSFER_CONTEXT, from-internal-xfer-dest, which dials the target through a Local channel. A Local channel take...
Cap llama-server's prompt cache below the memory limit, so it stops being OOMKilled
llama-server's host-RAM prompt cache defaults to 8192 MiB (--cache-ram), the container's whole limit, so it grew until the kernel killed ollama: on 2026-09-28 i...
The workstation probe shows its results and ends
Run live against an awake AI workstation, the probe printed the descriptor and then nothing: kubectl debug -i (1.36) returned as soon as it had sent the probe t...
The meet bridge writes the dial status at hangup, so answered calls record it
Once Jigasi answers, Dial never returns (no 'g' option): either side hanging up ends the channel, so the **************** after Dial ran only for calls that wer...
A call joins its room as jigasi@auth.meet.Meet, not anonymously; config changes restart it
Each call joins its room over its own XMPP login. Without ENABLE_AUTH the image makes that login ANONYMOUS on meet.Meet, which Prosody's JWT host refuses ("SASL...
Probe the XMPP session, not /about/health, which a SIP gateway answers 500
Jigasi 1.1.369's Health.check throws "GW not ready." whenever it has a SIP gateway and no HEALTH_CHECK_SIP_URI, registered or not, so the health probes restarte...
An AI's scroll scrolls Chrome
computer-use-linux 0.7.1 aims a scroll with ydotool, whose absolute move passes through 0,0 (the panel) on its way to the target, and sends the whole scroll as ...
Registry prune waits for CI before garbage-collect
garbage-collect sweeps every blob the tags didn't reference when it marked, so a push that lands mid-sweep keeps its tag and loses its manifest. On 2026-09-28 t...
The load verdict's budget is the amended acceptance bound, 9.0 s
Task 18 A-3 asked for p95 <= 4.0 s with 4 realtime streams and a batch. k3m1's Haswell CPU cannot reach it with faster-whisper-small: with the language given (S...
Measure realtime load as SP13 calls it, and give each transcription 8 threads
The load probe's realtime sessions sent no language, so every turn paid Whisper's language detection (9.4 s alone at 3 threads on k3m1), a call shape no product...
Run llama-server with one thread per core of its CPU limit
Ollama starts llama-server without --threads, so it counted k3m1's 24 physical cores and ran 24 compute threads inside the container's 8-core quota. The contain...
Ydotool's pointer lands where the AI aims it
ydotool 1.0 aims an absolute move by pushing its relative device to 0,0 and then by (x, y), and libinput's default adaptive profile doubled that second push. co...
The gateway may frame its own resize sensor
Guacamole 1.5.5 notices a resize through an <object> that loads its own **************** The gateway's X-Frame-Options: DENY blocked it (ERR_BLOCKED_BY_RESPONSE...
An AI workstation shows its organization's wallpaper
KubeVirt's console runs Xorg's VESA driver (RandR 1.1, no named output), so xfdesktop keys the AI desktop's backdrop by index, monitor0, and never read the moni...
The AI's console X server takes the AI's clicks
LightDM's Xorg on :0, the AI workstation's desktop, adds udev input devices only through an input driver, and the golden image had none. It logged 'No input dri...
Register with FreePBX under the names the image reads; one owner; health probes; host-only ingress
Size the reranker's TEI thread pools to its 1-CPU limit
The reranker runs embedding-model's TEI 1.9.4 under the same 1-CPU limit, so it takes the same sizing (KamoAI SP00 final review I-1): one tokenization worker, a...
Record live kontak.kamocrm.com, desktop-1 sizing and the Docs host list
A .forgejo/ change makes the next deploy a full reconcile, which applies every manifest over live state. Four had drifted from live by hand or by another repo's...
Size every TEI thread pool to the 1-CPU limit, and let probes outlast a batch
TEI 1.5's ONNX backend starts one inference thread per physical host core (24 on k1m1) and reads no thread setting, so under the pod's 1-CPU quota the CFS throt...
Transcribe realtime turns over the pod's own loopback
Speaches 0.9.0-rc.3 posts every realtime turn to /v1/audio/transcriptions. With LOOPBACK_HOST_URL unset it makes that call in-process through httpx's ASGITransp...
Reserve 768M for the OS so --memory 3G fits in the 4Gi pod
Seastar keeps max(1.5 GiB, 7%) of the cgroup limit plus ~94 MiB per shard back for the OS and refuses to start when --memory does not fit in what is left. With ...
X keeps counting through a sleep, so every resumed computer went straight back to sleep
A member opening their computer got a black tab. The launch was fine, the wake was fine, the resume was fine — and then the machine powered itself off before th...
A cloud session asks for its task in the terminal
description") unless the account has the interactive cloud backend, which this one does not; given one, it creates the session, prints its task for the cloud se...
ASCII in finalize.sh's log lines — jibri's logger printed the em dash as ???
Xrdp was told its session backend was itself, so nothing ever rendered
Opening a computer gave a black screen that never became a desktop. The cause is one unanchored sed in this file: sed -i 's/^port=.*/port=3389/' /etc/xrdp/...
Recording had never worked — jicofo never looked for jibri, and jibri never arrived
Four separate faults, each enough on its own to stop every recording: - jicofo had no ENABLE_RECORDING, so the image wrote no `jibri {}` block: it never watc...
Recreate, and probe the XMPP session — a surge left meetings with no focus for 18h
Every jicofo logs in to prosody as the same full JID, focus@auth.meet.Meet/focus. The maxSurge rollout from 9da0998 runs two of them at once, and prosody answer...
Two threads rebuilt the initrd at once and destroyed the machine
A computer provisioned, came up, and panicked on its next boot: Run /init as init process Failed to execute /init (error -2) Kernel panic - not syn...
The restart could land mid-initramfs and leave a computer unbootable
A brand new computer provisioned, came up, reported AWAKE, and ninety seconds later panicked: Kernel panic - not syncing: No working init found. I caused ...
A resumed computer went straight back to sleep
Thirty seconds after the first real resume, from the machine's own log: kamo-hc-agent: idle 13202s >= 1800s — going to sleep PM: hibernation: hibernati...
Kamouniverse.com redirect was losing to the www.* catch-all
www.kamouniverse.com was served by the kamocrm.com marketing app instead of redirecting: kamo-marketing-route's HostRegexp(\`^www[.].+\$\`) at priority 24 match...
The login greeter was counting as a member using the computer
rdp_sessions() collected every Xorg process with a display argument. The image installs lightdm and lightdm-gtk-greeter, so lightdm runs its own Xorg on :0 on e...
A stale resume_offset would hibernate into the wrong place
resume_configured() checked that `resume=` and `resume_offset=` were present on the running kernel's command line. Present, not correct — and the difference is ...
Root could not read a member's idle time, so nothing ever slept
The first computer reported its idle time as 13.9, then 29.2, then 14.4 seconds over three minutes while nobody was touching it. That is not a person — it is a ...
The swap file was smaller than the check that measures it
With the restart landing, the first computer came up with resumeConfigured true and stopped on the next precondition — which turns out to be unsatisfiable by co...
A computer could never restart into being able to sleep
The first real Hosted Computer reached AWAKE and then stayed awake forever. Hibernation reported resumeConfigured: false, IdleWatcher only arms when readiness i...
Like what you see shipping?
All of it arrives in your workspace on its own. Start on the free plan and read this page again in a month.