Fibmesh Outbound / CLI & API

Inspect the profile.
Verify the route.

The development CLI lists egress profiles. The API defines separate creation, enable and disable operations. A recorded change still needs a working exit and a real traffic test.

Read the implemented command Development CLI & API reference · Confirm platform and deployment support during assisted setup.

Available with assisted setup. Confirm the supported platform, source address and full-tunnel policy before connecting.

01 / Development command

List the records.
Know the boundary.

fibmesh egress list is the only implemented egress command, with text or JSON output. Creation, enable and disable mutation commands are not implemented. The login command is also a pending implementation.

These are development tools and contract references. They are not a production installation or authenticated signup guide.

Implemented / read onlyTwo output formats. The same profile list.
fibmesh egress list --format text
fibmesh egress list --format json

Requests GET /v1/egress-profiles

Profile list
Administrative records. No tunnel or source-address test.
fibmesh status
fibmesh diagnostics
Local identity and cached-policy information. No live route test.

02 / API operations

Read the operation.
Then read its effect.

The contract defines four authenticated user operations. HTTP success and administrative state each need careful interpretation.

On a narrow screen, scroll the table sideways to read the full endpoints and effects.

Current egress-profile contract
Operation / endpointResponseAdministrative effect
ListGET /v1/egress-profilesProfile recordsReturns the profiles for the authorised organisation.
CreatePOST /v1/egress-profiles201 · CreatedCreates an initially disabled record. It does not provision an exit gateway or allocate a public source.
EnablePOST /v1/egress-profiles/{id}/enable202 · AcceptedAccepts the state change and disables other profiles in the organisation. Verify applied policy and delivery separately.
DisablePOST /v1/egress-profiles/{id}/disable202 · AcceptedAccepts the state change. Verify native route restoration on the actual platform.

What the record says

A saved profile
is one observation.

Creation does not provision an exit. gateway_node_id and assigned_public_ip may be empty. An active or enabled record describes administrative state.

What delivery requires

A real request
is another.

Confirm an assigned, working exit, the endpoint’s applied policy and execution result, then test the destination from the actual application. Verify its source, DNS and each IP family.

Neither a record nor HTTP acceptance proves that internet traffic reaches a destination.

03 / Names in the contract

Four mode identifiers.
One full-tunnel intent.

These are API IDs within Outbound, not separate public products. Match the requested mode to the agreed service and assigned delivery.

Required
name + mode
Optional
region
Not accepted by create
Application selector, destination list or per-device exit binding.
secure_internet
Internet routing through a Fibmesh exit

Confirm the supported platform, assigned exit and actual traffic behaviour.

regular_vpn
Regular exit routing

No fixed or exclusive source address is implied.

stable_ip
A recognised outgoing source

Agree the assigned address, continuity and retention terms.

dedicated_ip
Exclusive outgoing identity

Confirm allocation, exclusivity and commercial scope.

Routing scope

The current policy builder requests full-tunnel routing for every mode, including 0.0.0.0/0 and ::/0. That does not certify working IPv6 delivery or eliminate operating-system exceptions. Verify both families and effective routes on the supported platform.

Organisation scope

Enabling selects one active exit for the organisation and disables its other profiles. These operations do not provide per-user region switching or simultaneous per-device exits.

04 / Automation & recovery

Bound the wait.
Keep a way back.

Start with the organisation and the intended change. Establish supported authentication and delivery before depending on an automated workflow.

  1. Read the current state

    Identify the organisation, existing profiles and active exit before writing. Enabling selects one organisation-wide profile; avoid replacing another team’s exit unexpectedly.

  2. Reconcile the response

    Record the returned profile ID. If a create call times out, inspect existing records before retrying: these operations do not establish an idempotency-key contract.

  3. Wait for applied delivery

    Confirm the assigned gateway and source, then the affected endpoint’s new policy and execution result. Use a bounded wait with a clear failure outcome; never poll indefinitely for assumed success.

  4. Test, disable and recover

    Test the actual destination’s observed source, application response, DNS and IPv4/IPv6 behaviour. Exercise the agreed disable and recovery procedure, then verify ordinary routing with a fresh connection.

Keep credentials at the right boundary

Keep endpoint private keys local and redact tokens, keys and other secrets from logs. Never put hosted database credentials or service-role keys in a customer CLI or browser.

Use the deployment verification guide →

For a supported deployment

Bring the workload.
Agree the operating path.

Share the platform, intended exit, source-address needs and automation requirements. Confirm the supported environment, access and recovery procedure with Fibmesh during assisted setup.