TCP monitors
Check that a TCP port accepts connections — for databases, brokers and anything that isn't HTTP.
A TCP monitor opens a connection to a host and port and reports success if the connection is accepted within the timeout.
Configuration
| Setting | Allowed values | Required |
|---|---|---|
host |
hostname or IP, 1–255 characters | yes |
port |
1–65535 | yes |
Everything else — interval, timeout, retries, regions, confirmations — is shared with all monitors.
What it does and does not tell you
A successful TCP check proves the port is open and something accepted the connection. It does not prove the service behind it is healthy: a database that has run out of connections, or is stuck replaying a log, will usually still complete a TCP handshake.
Use TCP when there is no richer check available. When the service speaks HTTP, an HTTP monitor with an assertion tells you far more.
Common uses
| Service | Typical port |
|---|---|
| PostgreSQL | 5432 |
| MySQL / MariaDB | 3306 |
| Redis / Valkey | 6379 |
| SMTP | 25, 465, 587 |
| SSH | 22 |