Struct Manager 

pub struct Manager { /* private fields */ }

The ingress: keeps Caddy’s config in line with the stacks.

Implementations

impl Manager

pub fn set_extras(&self, extras: Vec<ExtraRoute>)

Replace the extra routes: one caller (the daemon’s workspaces) owns them all. Caddy is reloaded only when they changed.

pub fn check_extra(&self, r: &Route) -> Result<()>

Check a new extra route before it is kept: the org’s allowlist and provider, and no other claim on its name.

pub fn extra_status(&self, r: &Route) -> Vec<DomainStatus>

An extra route’s state, as stack_status reports a domain’s.

impl Manager

pub fn status(&self, orgs: Option<&[OrgId]>) -> Value

Everything the ingress_status tool shows, for the given orgs (all when None).

impl Manager

pub fn new( cfg: IngressConfig, client: Client, secrets: Arc<Secrets>, state_dir: &Path, ) -> Result<Arc<Manager>>

Create the manager. Caddy starts with Manager::start, once the controller exists.

pub fn start(self: &Arc<Self>, ctl: Controller) -> Result<()>

Get Caddy (downloading the pinned release if needed), start it, and start the applier.

pub fn public_ip(&self) -> Option<IpAddr>

The address generated (host: auto) names point at.

pub fn shutdown(&self)

pub fn check(&self, def: &StackDef) -> Result<()>

Check a stack about to be deployed: its domains must be allowed for its org and free of other orgs’ (and other services’) claims.

Trait Implementations

impl Observer for Manager

fn rotation(&self, stack: &str, service: &str, ips: &[IpAddr])

The in-rotation replicas of a service changed. stack is qualified.

fn drain(&self, stack: &str, service: &str, ip: IpAddr, timeout: Duration)

Wait (at most timeout) until requests to ip have drained from the observer’s proxies, before the replica is stopped.

fn stacks_changed(&self, defs: Vec<Arc<StackDef>>)

The deployed definitions changed (deploy, scale, rollback, removal).

fn domains(&self, stack: &str, service: &str) -> Vec<DomainStatus>

A service’s domains for its status.