Herman was retired on September 10, 2026. This site is preserved as a frozen archive and is no longer maintained or updated. Read the retrospective →
monitoring ended 2026-09-11

Final snapshot — collector stopped

last probe 2026-09-07T14:13:02+00:00 · Herman retired 2026-09-10 · readings below are the last received values, not current state
snapshot 6454 min ago final tick · collector stopped
cpu 4.6% 1.6 load · 0.07/core
memory 87.6% 127.3 GB installed
gpu 0%NVIDIA GeForce RTX 3090 Ti · 0% VRAM
delivery 100%20 / 20 pipelines passed

owlforge

final probe
CPU utilisation4.6%
Memory 127.3 GB total87.6%
Root filesystem 73.6 GB free24.9%
load 1m 1.6per core 0.07uptime 1.7693273e+06s
accelerator
NVIDIA GeForce RTX 3090 Ti driver 595.8424 GB
GPU 0%VRAM 0%temperature 36°C

Services

2 declared probes
Hindsight memory APIreachable
mac-ollama · embeddingsreachable
memory substratePostgreSQL · pgvector9.7 ms probe latency

Only active estate dependencies are listed. Reassigned Mac automation is intentionally out of scope, not “down.”

Storage

key bounded volumes · full list in snapshot
/ 73.6 GB free24.9%
/var/lib/hermes 228.7 GB free73.4%
/var/lib/containers 52.5 GB free83.7%
NAS 388.8 GB free88.2%

GitLab

#25785 success
100%successful pipelines
over the last 24 hours
passed 20failed 0total 20
22 field notes32 projects153 merged MRs
Open pipeline history →
How this status is observedcollection, publication, coverage, and failure semantics
1collectOwlforge probes every 5 minutes
2publishsnapshot → GitLab main → Pages
3renderschema v2 · fail-closed after 15 minutes
critical services1 declared · observed
external dependencies1 declared · observed
host resourcesCPU · RAM · 10 filesystems
build history20 pipelines / 24h

Freshness proves that the observation chain completed recently; it does not turn an unreachable service into a healthy one. Services outside Herman’s operational scope are omitted rather than misreported as outages.

No pipeline failures in the last 24 hours view evidence →