Skip to content
How to install

Sites

Create, change and remove sites. Everything that touches a container is asynchronous: you get a job back and poll it.

Reading needs Read only and changes need Manage, unless an endpoint says otherwise. How to read this page.

List every site with its status, health probe, PHP release and WordPress snapshot.

  • Level: Read only
  • Returns: {items: SiteSummary[]}

Create a site: WordPress, a PHP site with no WordPress in it, or a container app.

  • Level: Manage
  • Notes: Job
  • Input: {title, slug?, serverId?, domainMode: "dev"|"custom", domains?, phpVersion?, app?: "wordpress"|"php"} and, for WordPress: {locale?, adminUser, adminEmail?, adminPassword?, discourageSearchEngines?, plugins?: {catalogIds?, extraWporgSlugs?}} - adminEmail left out uses the defaultAdminEmail setting (400 when that is empty); catalogIds left out installs the catalog's default plugins, [] none of them. For a PHP site (app "php"): {database?: true for a database of its own, webRoot?: a folder in the site, such as public}, and none of WordPress's fields (400 naming each). For a container app (app "container"): {image: {reference, registryUser?, registryToken?} or {uploadId}, port?, env?, routes? (left out, its dev hostname over https), mounts? (an app made with mounts is left stopped: fill them, then start it), dataPaths?, extraHosts?, database?, hardening?, health?, updates?}, as PATCH /api/sites/:slug/app takes them; no domains or phpVersion
  • Returns: job.result: {url, adminPassword} (WordPress), {url} (PHP, container app); 409 for a PHP site on a server whose image is from before PHP sites; 403 below Full for a container app route on another site's domain

One site in full, including its container state, monitoring snapshot and the job it is busy with.

  • Level: Read only
  • Returns: SiteDetail; a container app's failed deploys carry their logTail only from Manage up

Delete a site, its database and its files.

  • Level: Full. Removes the site itself, with its database and files
  • Notes: Job · Destructive
  • Input: ?finalBackup=true|false (default true)&deleteBackups=true|false (default false; with a final backup, that one is kept)

Start the container.

  • Level: Manage
  • Notes: Job

Stop the container.

  • Level: Manage
  • Notes: Job · Destructive

Restart the container.

  • Level: Manage
  • Notes: Job

Switch PHP version, rolling back if the site stops answering.

  • Level: Manage
  • Notes: Job
  • Input: {phpVersion}

A PHP site’s web root: the folder visitors reach, in force at once.

  • Level: Manage
  • Input: {webRoot} - a folder inside the site's, such as public; '' is the site's folder
  • Returns: SiteDetail; 409 for a WordPress site, or with Apache's words when it refuses the file

A PHP site’s cron lines, each with its schedule in words and its last run.

  • Level: Manage. The lines are commands, and the output of the last runs is shown with them
  • Returns: {lines: SiteCronLineDto[]}; 409 for a WordPress site

Set a PHP site’s cron lines: up to 10 commands run in its folder on a schedule, outside the job queue.

  • Level: Manage
  • Notes: Destructive
  • Input: {lines: [{id?, expr: "*/5 * * * *", command, enabled?, timeoutMin?: 1-60, default 5}]} - every line as it is to be: one with an id keeps its last run, one left out is deleted; five fields on the panel's clock, every minute allowed
  • Returns: {lines: SiteCronLineDto[]}; 400 for a schedule that would never run: one croner refuses, or that no date matches

Run one cron line now, as the next minute would; its result lands on the line.

  • Level: Manage
  • Notes: Destructive
  • Returns: {line: SiteCronLineDto}; 409 while it runs, while the site is stopped or busy, or with 2 already running on its server

Send a test message as the site’s own code would: wp_mail(), or PHP’s mail() on a PHP site.

  • Level: Manage
  • Input: {to}
  • Returns: {accepted, detail}

Move a dev site onto its real domains with no downtime.

  • Level: Manage
  • Notes: Job · Destructive
  • Input: {domains: [primary, ...aliases], keepDevAlias?: true, allowSearchEngines?: true, manageDns?: false (true needs Full)}

Edit the domains a live site answers on.

  • Level: Manage
  • Notes: Job · Destructive
  • Input: {domains}

Move the site to another server; the old one forwards until DNS catches up.

  • Level: Full. Crosses servers and leaves a parked copy behind
  • Notes: Job
  • Input: {targetServerId, quiesce?: "maintenance"|"stop"|"none"}
  • Returns: The job; 409 for a container app with plain http routes when the target's Traefik is from before them

Tear down the parked source copy now instead of waiting for DNS verification.

  • Level: Full. Ends a move, which is Full: tears down the parked copy
  • Notes: Job · Destructive

Re-apply the isolation policy (network, capabilities, limits, mail credential); retries a failed PHP-FPM move.

  • Level: Manage
  • Notes: Job

The same for every site, one job each.

  • Level: Manage
  • Notes: Job
  • Returns: 202 {jobs}

Set the site’s page cache, how long it keeps pages, its rules, and its object cache.

  • Level: Manage
  • Notes: Job
  • Input: {page?: default|on|off, object?: boolean, keepSeconds?: number|null, off?: string[], own?: rule[]} - each left out = unchanged. keepSeconds: at most 36000, null follows Settings. off: ids of built-in groups and rules to switch off; the whole list. own: the site's own rules, {kind: skip|copy|keep|ignore, match: path|cookie|query|agent|postType, value, seconds?}; the whole list
  • Returns: {site: SiteDetail, job} - job: the site.objectCache when the object cache changes; 200 and null otherwise

Empty the site’s page cache, and Redis when it runs.

  • Level: Manage
  • Returns: {cleared: boolean}

Ask for a page as a visitor with no cookies would, and say what the page cache did with it.

  • Level: Manage
  • Input: {path} - starting with /
  • Returns: {state: cached|stale|skipped|not-stored|off, reason, status} - reason: the rule or the page's own words

Set the site container’s own CPU, memory and process limits, or put them back on the panel’s settings.

  • Level: Full. More for one site is less for the other sites on its server
  • Notes: Job
  • Input: {cpuLimit?, memoryLimitMb?, pidsLimit?} - each a number, or null for Settings' value; left out = unchanged. CPU in cores; 0 = uncapped, for CPU and processes; a process limit is at least 64
  • Returns: {site: SiteDetail, job} - job: the server.applySiteLimits that changes the container in place (lifting a CPU cap recreates it); 200 and null when nothing it runs with changed

Include or exclude this site from the scheduled backup run.

  • Level: Full. Switching backups off removes a safety net
  • Notes: Destructive
  • Input: {enabled}
  • Returns: SiteDetail (sync)

Stop or resume copying this site’s backups offsite.

  • Level: Full. Switching offsite copies off removes a safety net
  • Notes: Destructive
  • Input: {enabled}
  • Returns: SiteDetail (sync)

Set after how many days this site’s backups are deleted, or put it back on the panel’s setting.

  • Level: Full. A shorter keep period deletes backups, a safety net
  • Notes: Destructive
  • Input: {days: number | null} (null = Settings, 0 = never)
  • Returns: SiteDetail (sync)

PUT /api/sites/:slug/backup-before-updates

Section titled “PUT /api/sites/:slug/backup-before-updates”

Set whether an update backs this site up first, or put it back on the panel’s setting.

  • Level: Full. Never backing up before an update removes a safety net
  • Notes: Destructive
  • Input: {value: "always" | "ask" | "never" | null} (null = Settings)
  • Returns: SiteDetail (sync)

Stop or resume the relay accepting this site’s mail.

  • Level: Full. Lifting a suspension undoes the spam guard
  • Notes: Destructive
  • Input: {suspended, reason?}
  • Returns: SiteDetail (sync)

Visitor statistics: visitors, page views, top pages, referrers, countries, crawlers.

  • Level: Read only
  • Input: ?days=1..365 (default 30)
  • Returns: SiteTrafficDto