bluejay/bluejay-infra
1
0
Fork 0
Code Issues Pull Requests 14 Actions Packages Projects Releases Wiki Activity

Fix edge1 Ollama IP (.15->.17), add monitoring ingress, add init container

Browse Source
This commit is contained in:
Claude Code
2026-04-08 17:30:22 +00:00
parent f3919cf728
commit c9f07108bd
1 changed files with 368 additions and 325 deletions
Download Patch File Download Diff File Expand all files Collapse all files

57
apps/agent-zero/agent-zero.yaml
View File

@@ -2,7 +2,7 @@
# Agent Zero AI Stack — NUC Deployment (RKE2 Bare-Metal)
# =============================================================================
# Deploys: AgentZero (agent UI) on RKE2 cluster
# Ollama: edge1 Pi 5 at 10.0.57.15:11434 (qwen2.5-coder:7b, CPU)
# Ollama: edge1 Pi 5 at 10.0.57.17:11434 (qwen2.5-coder:7b, CPU)
# Target: RKE2 bare-metal cluster, namespace: agent-zero
#
# Differences from LOCAL (WSL K3s):
@@ -99,7 +99,7 @@ metadata:
app: agent-zero
annotations:
agent-zero/deployment: "nuc"
agent-zero/ollama: "edge1 Pi 5 (10.0.57.15:11434)"
agent-zero/ollama: "edge1 Pi 5 (10.0.57.17:11434)"
spec:
replicas: 1
selector:
@@ -113,6 +113,21 @@ spec:
app: agent-zero
spec:
serviceAccountName: agent-zero
initContainers:
# Wait for edge1 Pi 5 Ollama to be reachable before starting Agent Zero.
# Without this, the FAISS memory system crashes on embed_query()
# and Agent Zero enters a broken state on every message.
- name: wait-for-ollama
image: busybox:1.37
command: ["sh", "-c"]
args:
- |
echo "Waiting for Ollama at edge1 (10.0.57.17:11434)..."
until wget -qO- --timeout=2 http://10.0.57.17:11434/api/tags >/dev/null 2>&1; do
echo "edge1 Ollama not ready, retrying in 5s..."
sleep 5
done
echo "edge1 Ollama is reachable!"
containers:
- name: agent-zero
image: agent0ai/agent-zero:latest
@@ -141,7 +156,7 @@ spec:
- name: A0_SET_chat_model_name
value: "qwen2.5-coder:7b"
- name: A0_SET_chat_model_api_base
value: "http://10.0.57.15:11434"
value: "http://10.0.57.17:11434"
- name: A0_SET_chat_model_ctx_length
value: "32768"
- name: A0_SET_chat_model_kwargs
@@ -152,7 +167,7 @@ spec:
- name: A0_SET_util_model_name
value: "qwen2.5-coder:7b"
- name: A0_SET_util_model_api_base
value: "http://10.0.57.15:11434"
value: "http://10.0.57.17:11434"
- name: A0_SET_util_model_kwargs
value: '{"num_ctx": 8192}'
# Embedding model — nomic on edge1 (if installed, fallback to none)
@@ -161,14 +176,14 @@ spec:
- name: A0_SET_embed_model_name
value: "nomic-embed-text"
- name: A0_SET_embed_model_api_base
value: "http://10.0.57.15:11434"
value: "http://10.0.57.17:11434"
# Browser model — disabled (no vision model on Pi)
- name: A0_SET_browser_model_provider
value: "ollama"
- name: A0_SET_browser_model_name
value: "qwen2.5-coder:7b"
- name: A0_SET_browser_model_api_base
value: "http://10.0.57.15:11434"
value: "http://10.0.57.17:11434"
- name: A0_SET_browser_model_vision
value: "false"
# Agent profile
@@ -196,6 +211,27 @@ spec:
mountPath: /a0/work
- name: knowledge
mountPath: /a0/knowledge/custom/main
startupProbe:
httpGet:
path: /
port: 80
initialDelaySeconds: 15
periodSeconds: 10
failureThreshold: 18
livenessProbe:
httpGet:
path: /
port: 80
periodSeconds: 30
failureThreshold: 3
readinessProbe:
exec:
command:
- /bin/bash
- -c
- "curl -sf http://localhost:80/ > /dev/null && curl -sf --connect-timeout 3 http://10.0.57.17:11434/api/tags > /dev/null"
periodSeconds: 30
failureThreshold: 2
resources:
requests:
memory: "2Gi"
@@ -292,6 +328,13 @@ spec:
kubernetes.io/metadata.name: traefik-system
ports:
- port: 80
# Allow from monitoring (blackbox probe)
- from:
- namespaceSelector:
matchLabels:
kubernetes.io/metadata.name: monitoring
ports:
- port: 80
egress:
# DNS
- to:
@@ -306,7 +349,7 @@ spec:
# Ollama on edge1
- to:
- ipBlock:
cidr: 10.0.57.15/32
cidr: 10.0.57.17/32
ports:
- port: 11434
# K8s API