Skip to content

Integrations

← Back to docs index

Each server carries its integrations under the Settings group: Discord, a public Status Page, one-click Invites, and the optional Chatbot tab (admins only). This page covers Discord; invites & status page covers those two, and the chatbot has its own page.

The panel forwards server events to a Discord channel through a webhook. No bot, no OAuth.

  1. In Discord: Server Settings → Integrations → Webhooks → New Webhook. Pick the channel and copy the webhook URL.
  2. In the panel, open the server’s Settings → Discord tab, paste the URL into Webhook URL, turn the card on, and Save. The URL is stored encrypted (in the same at-rest store as your API keys).
  3. Use Send Test to confirm it works.

Discord integration

To replace a stored URL, paste a new one and save; leave the field blank to keep the current one. The stored URL is shown masked (the token is hidden).

Notifications are grouped into toggleable categories. All are on by default.

Category Events
Start / stop Server started, server stopped
Crashes Crash, crash loop detected, new crash report
Backups Backup created, backup restored
Updates Update applied, update rolled back, update failed
Kicks / bans Player kicked, player banned
Alerts The things that quietly leave a server broken (see below)

Alerts is the “something needs a human” channel, for conditions that used to be visible only in the panel:

Event Meaning
oom The container was stopped for running out of memory.
unhealthy The process is up but the mc-health probe is failing: a live but dead server.
startup-stalled Still starting after ~10 min with no “Done” in the logs (with a diagnosis of what to fix, where one is known).
stop-failed A graceful stop didn’t take effect and the container is still running.
schedule-failed A scheduled restart / backup / command failed.
quota-exceeded Strict-mode disk quota tripped and the server was stopped.
offline-after-restart The server was up before the panel restarted, isn’t now, and won’t be auto-started.
port-change-failed A port change failed and the server could not be put back on its old port, so it is down.
auto-restart-gave-up A crashed server was not restarted because another action kept it busy for too long.

A background bridge polls the event log every 15 seconds and forwards mapped rows. If a webhook POST fails transiently, that row is retried on later polls rather than dropped; after a few failed attempts the row is skipped, and delivery failures are logged at most once per hour per server. After a panel restart the bridge sends the events it had not delivered yet, but never more than two hours back; on its very first run it starts from the newest event.