One login. Every site. DHCP stays there. Fleet is a manager of managers — not a copy of cluster data. Each Lattice cluster keeps its own database and grants its own leases. You search a MAC, open the live console, and leave.
For people who run Lattice at more than one site — a service provider with mixed small and large customers, or a company with more than one datacenter. Community clusters and paid clusters sit in the same Fleet.
The data stays on the cluster
Each site keeps its own database, its own leases, and its own license file. Fleet does not replicate that catalog and is not a backup. Opening a cluster in Fleet is the live console — through a tunnel — not a second copy.
The site does not open inbound 443
An agent on the cluster calls out to Fleet over HTTPS. Relays still point at that site. Fleet never serves DHCP.
A second license, counted by clusters
Community is two clusters with no Fleet file. Paid tiers: 10, 25, 50, 200, 500. Same names and list prices as cluster Lattice. Mix Community sites and paid sites in one Fleet.
Manager of managers
One login. Each cluster keeps its own database. You search a MAC, open the live console, and leave. Nothing is copied to the manager. An isolated site is another register — not another system.
Scale
One console, sites stay separate
No
Windows
Partial
Stork
Yes
Fleet
No
Grid
Add a site, same login
No
Windows
Partial
Stork
Yes
Fleet
No
Grid
Privacy
No replica of leases
Partial
Windows
Yes
Stork
Yes
Fleet
No
Grid
Security
No inbound 443
No
Windows
Yes
Stork
Yes
Fleet
No
Grid
Audited break-glass
No
Windows
No
Stork
Yes
Fleet
Partial
Grid
Operate
Search a MAC across sites
Partial
Windows
Partial
Stork
Yes
Fleet
Partial
Grid
Occupancy across sites, no replica of leases
No
Windows
Partial
Stork
Yes
Fleet
No
Grid
Edit from this console
No
Windows
Partial
Stork
Yes
Fleet
Partial
Grid
One alert inbox
No
Windows
Partial
Stork
Yes
Fleet
Yes
Grid
Compare Windows, Stork, Fleet, and Grid
Outcome
Windows
Stork
Fleet
Grid
Scale
One console, sites stay separate
No
Partial
Yes
No
Add a site, same login
No
Partial
Yes
No
Privacy
No replica of leases
Partial
Yes
Yes
No
Security
No inbound 443
No
Yes
Yes
No
Audited break-glass
No
No
Yes
Partial
Operate
Search a MAC across sites
Partial
Partial
Yes
Partial
Occupancy across sites, no replica of leases
No
Partial
Yes
No
Edit from this console
No
Partial
Yes
Partial
One alert inbox
No
Partial
Yes
Yes
Yes
Partial
No
In the Fleet console
Overview
Health of every cluster: version, license, agent connected, hottest subnets. Click a tile and you are in that site.
The live cluster console
Addresses, leases, Work, DNS, tenants, tokens — the same screens as logging into the site. Switch cluster and stay on the same page.
Inbox
Alerts from every cluster in one list. Click the title, land on the lease or subnet. Ack or snooze. The grant already happened on the site.
Search and History
Paste a MAC, see every site that has a hold, open the live cluster. Nothing is copied. History fans out live occupancy the same way — as-of needs that cluster’s lease table. Paste includes switch ports.
What Fleet is not
Not a replica of cluster data. Not a backup of leases or IPAM.
Not a second DHCP path. Relays never point at Fleet.
Not a hosted service. You run lattice-fleet.
Not a cluster feature flag. Separate binary, separate file.
Not a way to run two isolated address plans on one dhcpd. That is two clusters; Fleet is how you operate both from one login.