Uptimeify Docs
MonitoringMonitoring types

TCP Port Monitor

The TCP Port Monitor checks whether a raw TCP port is open and accepting connections. It is commonly used for databases, caches, message queues, and other non-HTTP services that don't speak a protocol we have a dedicated monitor for.

How it works

  • We attempt to open a TCP connection to your hostname:port.
  • If the connection cannot be established or times out, the check fails.
  • If you configure Expect Banner, we also verify that the raw bytes received right after the handshake contain that substring (e.g. +PONG for Redis, 220 for an SMTP banner) — otherwise the check fails even though the port is open.

Configuration

Basic settings

  • Hostname: Hostname or IP address (no protocol, no path).
  • Port: Required. Unlike other monitor types, there is no default port — you must specify the exact port to connect to (e.g. 6379 for Redis, 5432 for PostgreSQL, 27017 for MongoDB).
  • Expect Banner: Optional substring the worker must find in the bytes received immediately after the TCP handshake.
  • Check interval and Timeout.
  • Status: active, maintenance, disabled.

Monitoring locations

TCP checks run from our monitoring locations. Use Allowed Check Countries to limit where checks run from.

No credentials

TCP Port monitors carry no credentials or authentication — they only verify that the port accepts connections (and optionally that the initial banner matches). For services that require authentication before you can confirm they're healthy, use a protocol-specific monitor (e.g. SSH, SMTP, IMAP/POP) where one exists.

On this page