KV Cache Discovery
Tensormesh · LMCache
Monitor
Overview
Cache Browser
Analyze
Prompt Lookup
Utilization
MP Coordinator
Toggle Sidebar
Monitor
Overview
Overview
Fleet health, directory size and how much of the committed L2 quota the cache is actually using.
as of 2026-10-04 00:37:45 UTC
Refresh
MP servers
Bytes by tier
Per-tenant L2 usage
2/3 instances
|
3.2K keys
Instances healthy
2/3
1 degraded · 1 stream gap
Fleet keys
3,246
Unique cache keys tracked by the directory
Placements
4,204
1.30× copies per key
L1
519
L2
3,685
L2 usage vs quota
66%
66.3 GB of 100.0 GB committed · 20.8 GB unlimited
batch-eval over · team-search near
MP servers
3 registered · 2 healthy · 519 L1 keys
1 stream gap
Instance
Incarnation
L1 keys
Event stream
Browse keys
Degraded
mp-server-b1
10.132.0.21:8100
#2
113
gap detected
seq 17,089
Keys
Healthy
mp-server-a1
10.128.0.11:8100
#7
216
in sync
seq 19,042
Keys
Healthy
mp-server-a2
10.128.0.12:8100
#3
190
in sync
seq 71,423
Keys
Degraded
mp-server-b1
gap detected
Incarnation
#2
L1 keys
113
Browse keys
Healthy
mp-server-a1
in sync
Incarnation
#7
L1 keys
216
Browse keys
Healthy
mp-server-a2
in sync
Incarnation
#3
L1 keys
190
Browse keys
Bytes by tier
Fleet-wide placement footprint: L1 is the hot layer, L2 the capacity layer behind it.
92.11 GiB
across 4,204 placements
Fleet-wide cache bytes by tier
Tier
Placements
Bytes
Share of bytes
L1 · hot
519
11.06 GiB
12%
L2 · capacity
3,685
81.05 GiB
88%
Per-tenant L2 usage
Footprint against quota for every cache salt, largest first.
team-search
29.5 GB
/ 31.0 GB
95%
team-support
24.1 GB
/ 57.0 GB
42%
default
20.8 GB
· no quota
batch-eval
12.7 GB
/ 12.0 GB
106%
87.0 GB total · 3 of 4 tenants have a quota · 100.0 GB committed