Skip to content

Commit a73d29b

Browse files
committed
docs(FR-2603): translate model store heading and fix Confirm→Update button in ja/th, translate model serving occurrences in th
1 parent 4475424 commit a73d29b

12 files changed

Lines changed: 688 additions & 208 deletions

File tree

Lines changed: 85 additions & 0 deletions
Loading
Lines changed: 80 additions & 0 deletions
Loading

packages/backend.ai-webui-docs/src/en/model_serving/model_serving.md

Lines changed: 7 additions & 52 deletions
Original file line numberDiff line numberDiff line change
@@ -142,58 +142,13 @@ Fields without "(Required)" mark are optional.
142142
The health check system monitors individual model service containers and automatically
143143
manages traffic routing based on their health status.
144144

145-
```
146-
Container Created
147-
148-
149-
┌─────────────────────────────────┐
150-
│ Wait for initial_delay (60s) │ ← Model loading, GPU init, warmup
151-
│ Status: NOT_CHECKED │
152-
│ No health checks during this │
153-
└─────────────────────────────────┘
154-
155-
156-
Start Health Check Cycle
157-
158-
159-
┌─────────────────────────────────┐
160-
│ Every interval (10s): │
161-
│ HTTP GET → path ("/health") │
162-
└─────────────────────────────────┘
163-
164-
165-
Wait up to max_wait_time (15s)
166-
167-
┌──────────┴──────────┐
168-
▼ ▼
169-
Response Timeout/Error
170-
│ │
171-
▼ │
172-
Status == │
173-
expected? │
174-
│ │
175-
┌──┴──┐ │
176-
▼ ▼ │
177-
Y N │
178-
│ │ │
179-
│ └───────┬───────┘
180-
│ ▼
181-
│ Consecutive
182-
│ failures +1
183-
│ │
184-
▼ ▼
185-
HEALTHY Failures > max_retries?
186-
(reset │
187-
failures) ┌─────┴─────┐
188-
▼ ▼
189-
Yes No
190-
│ │
191-
▼ ▼
192-
UNHEALTHY Keep current
193-
(removed status
194-
from traffic
195-
internally)
196-
```
145+
**① AppProxy: Traffic Routing Control**
146+
147+
![](../images/health_check_app_proxy.svg)
148+
149+
**② Manager: Health State Management and Eviction**
150+
151+
![](../images/health_check_state_machine.svg)
197152

198153
:::note
199154
The internal health status (used for traffic routing) may not be immediately
Lines changed: 85 additions & 0 deletions
Loading

0 commit comments

Comments
 (0)