Communication security
Kubling uses a transport-independent securing mechanism, which works at communication level using SSL (or TLS). In other words, it is possible to secure native and PostgreSQL transport protocols using the same standard TLS configuration.
When a secure connection is established, the packets exchanged between clients and the Kubling instance are encrypted using TLS 1.3
, making the
data exchanged completely secure.
Enabling security
In each transport configuration, you can enable security by activating the transports.[protocol].secure: true
.
However, when enabled, also a server certificate must be specified, otherwise the TLS configuration will be ignored.
Generate Certificate
kdv
CLI provides a convenient command to generate certificates, that can be directly read by the engine, without having to convert formats.
kdv cert create -s /certs/server.ks -c /certs/client.ks -x myserverpass -p myclientpass -o
Command generates both client and server certificates. Client certificate is only required in case the client is configured to validate the server certificate.
When no password is provided for either the client or server certificates, kdv will automatically generate them and display the passwords in the console. Make sure to copy and store these passwords securely.
Configure transports
For security reasons, you may want to use a different certificate per transport, reason why their configurations are independent, as follows:
transports:
nativeProtocol:
bindAddress: "0.0.0.0"
portNumber: 35482
secure: true
sslConfig:
storeFilePath: "/certs/server-native.ks"
storePassword: "123abc"
pgProtocol:
bindAddress: "0.0.0.0"
portNumber: 35432
secure: true
sslConfig:
storeFilePath: "/certs/server-pg.ks"
storePassword: "456def"
Configure IP Whitelist and Blacklist v25.4+
Kubling supports IP-level whitelisting and blacklisting for inbound connections at the transport layer. This filtering occurs before messages enter the transport pipeline (but after the TLS negotiation, if configured), allowing Kubling to reject unwanted traffic as early and efficiently as possible.
Filters are applied per protocol (nativeProtocol
, pgProtocol
) and are defined using CIDR blocks, which can match single IPs or entire subnets.
Filtering is evaluated immediately when a remote client attempts to connect. If the IP is not allowed (or explicitly blocked), the connection is terminated without processing any payload.
Example
transports:
nativeProtocol:
...
ipFilter:
type: WHITELIST
cidrBlocks:
- 127.0.0.1 # localhost only
- 192.168.1.15 # single IP
- 10.10.0.0/16 # entire subnet
- 172.20.30.0/24 # class C block
pgProtocol:
...
ipFilter:
type: BLACKLIST
cidrBlocks:
- 192.168.1.1 # explicitly deny this IP
- 203.0.113.0/24 # deny this public test block
Notes
type
must be eitherWHITELIST
orBLACKLIST
(case-sensitive).- CIDR blocks can represent:
- A single IP (e.g.
192.168.1.15
) - A range/subnet (e.g.
10.10.0.0/16
)
- A single IP (e.g.
- Each protocol can define independent filters.
Use whitelists in secure environments where access should be strictly controlled, and blacklists when you only need to exclude known threats or test ranges.