Sessions - Mirage

The sessions store keeps each session’s shell state and the workspace’s own record:

Each session’s working directory, variables, functions, mount modes and profile rules. Shell options and the last exit status stay with the running instance.
Which sessions the workspace has, and which one is the default.

A workspace built again under the same id finds its sessions where it left them, and every instance of the workspace on the same store shares them.

`type`	Keeps them	Keys
`ram`	In the process’s memory, until it exits
`disk`	In files under `root`	`root`, default `~/.mirage/state`
`redis`	On Redis, under `key_prefix` and the workspace id	`url`, `key_prefix`, default `mirage:`
`s3`	In a bucket, under `key_prefix` and the workspace id	The S3 config fields, `key_prefix`, default `mirage/`

The group key is store.workspace. It is the only store that takes s3, so s3 goes on this key, never on store.type.

On disk

YAML
Python
TypeScript

workspace_id: agent
store:
  workspace:
    type: disk
    root: /var/lib/mirage

from mirage import Workspace
from mirage.workspace.store import DiskWorkspaceStateStore, RAMWorkspaceStateStore

store = RAMWorkspaceStateStore(
    workspace=DiskWorkspaceStateStore(root="/var/lib/mirage"),
)
ws = Workspace(mounts, workspace_id="agent", store=store, owns_store=True)

import { DiskWorkspaceStateStore, RAMWorkspaceStateStore, Workspace } from '@struktoai/mirage-node'

const store = new RAMWorkspaceStateStore({
  workspace: new DiskWorkspaceStateStore({ root: '/var/lib/mirage' }),
})
const ws = new Workspace(mounts, { workspaceId: 'agent', store, ownsStore: true })

On Redis

YAML
Python
TypeScript

workspace_id: agent
store:
  workspace:
    type: redis
    url: redis://redis.internal:6379/0

from mirage import Workspace
from mirage.workspace.store import RAMWorkspaceStateStore, RedisWorkspaceStateStore

store = RAMWorkspaceStateStore(
    workspace=RedisWorkspaceStateStore(url="redis://redis.internal:6379/0"),
)
ws = Workspace(mounts, workspace_id="agent", store=store, owns_store=True)

import { RAMWorkspaceStateStore, RedisWorkspaceStateStore, Workspace } from '@struktoai/mirage-node'

const store = new RAMWorkspaceStateStore({
  workspace: new RedisWorkspaceStateStore({ url: 'redis://redis.internal:6379/0' }),
})
const ws = new Workspace(mounts, { workspaceId: 'agent', store, ownsStore: true })

In an S3 bucket

The block takes the same fields as an S3 mount, so credentials and an endpoint_url for an S3-compatible store go here too.

YAML
Python
TypeScript

workspace_id: agent
store:
  workspace:
    type: s3
    bucket: agent-state
    region: us-east-1

from mirage import Workspace
from mirage.vfs.s3 import S3Config
from mirage.workspace.store import RAMWorkspaceStateStore, S3WorkspaceStateStore

store = RAMWorkspaceStateStore(
    workspace=S3WorkspaceStateStore(S3Config(bucket="agent-state", region="us-east-1")),
)
ws = Workspace(mounts, workspace_id="agent", store=store, owns_store=True)

import { RAMWorkspaceStateStore, S3WorkspaceStateStore, Workspace } from '@struktoai/mirage-node'

const store = new RAMWorkspaceStateStore({
  workspace: new S3WorkspaceStateStore({ bucket: 'agent-state', region: 'us-east-1' }),
})
const ws = new Workspace(mounts, { workspaceId: 'agent', store, ownsStore: true })