Network

See every blocked connection — created by an agent's ban or flagged by the automatic flood detector — with why it was blocked, how long is left, and a one-click release.

What a "connection" means

A connection is the visitor's internet address — the IP address their browser connects from. For IPv6 addresses, which change constantly by design, it is the household-sized block of addresses instead of a single one. Blocking a connection stops the trick every banned troublemaker tries first: opening a private window, or clearing their site data, to come back as a brand-new visitor.

A connection block is deliberately narrow:

  • It only stops new visitor sessions and new conversations from that connection on your widget. Anyone already mid-conversation on the same connection — a colleague on the office Wi-Fi, another customer behind the same mobile network — is never affected.
  • It lasts at most 24 hours, however long the ban that created it. Internet addresses get reassigned, so a block that outlived a day could land on a stranger.
  • A visitor turned away by it sees the usual quietly disabled chat input — no message about blocks or durations.

Where blocks come from

SourceHow it gets here
Agent banSomeone banned a visitor from a conversation with Also block their connection ticked — the default on a full ban. See Inbox Sidebar Actions. The block lasts as long as the ban, capped at 24 hours, and lifting the ban releases it.
AutomaticYaplet's flood detector noticed one connection creating an unusual number of new visitors or new conversations within 24 hours — the signature of a script, not a person.
The automatic detector is in watch-only mode for now. It records what it would have blocked — those rows carry the status Would have blocked — but blocks nothing on its own yet, and a banner on the tab says so. This lets its judgement be checked against real traffic before it is allowed to act. Connection blocks created by your team's bans work fully regardless.

The list

The tab shows every open block plus the last 30 days of history. Opening it needs the Security permission, like the rest of the Security page.

ColumnWhat it shows
ConnectionThe address that was blocked. IPv6 entries add a note with the block of addresses covered.
SourceAgent ban or Automatic.
StatusBlocked (in force now), Would have blocked (automatic, watch-only), Expired (ran out on its own), or Released (lifted by a person).
WhyThe ban reason for an agent ban; for the detector, the count that tripped it — for example 23 new chats from this connection in 24 hours.
CreatedWhen, and who created it.
ExpiresWhen it ends, with the hours left — or who released it.

Releasing a block

Release on a Blocked or Would have blocked row lifts the block immediately. Only the connection part is lifted: the ban on the visitor it came from stays in force — unban that from the conversation or from the Banned visitors tab. Anyone the connection block had turned away is let back in the next time their widget loads. Visitor opens the security profile of the visitor the block came from, where the block also appears in the "This visitor is blocked" banner.

If your own team suddenly can't start test conversations, look here first — a colleague banning a test visitor with the box ticked blocks the office connection for up to 24 hours, and one click releases it.