[−][src]Trait roughenough::config::ServerConfig

pub trait ServerConfig {
    fn interface(&self) -> &str;
    fn port(&self) -> u16;
    fn seed(&self) -> Vec<u8>;
    fn batch_size(&self) -> u8;
    fn status_interval(&self) -> Duration;
    fn kms_protection(&self) -> &KmsProtection;
    fn health_check_port(&self) -> Option<u16>;
    fn client_stats_enabled(&self) -> bool;
    fn fault_percentage(&self) -> u8;

    fn udp_socket_addr(&self) -> Result<SocketAddr, Error> { ... }
}

Specifies parameters needed to configure a Roughenough server.

Parameters labeled "Required" must always be provided and have no default value while those labeled "Optional" provide sane default values that can be overridden.

YAML Key	Environment Variable	Necessity	Description
`interface`	`ROUGHENOUGH_INTERFACE`	Required	IP address or interface name for listening to client requests
`port`	`ROUGHENOUGH_PORT`	Required	UDP port to listen for requests
`seed`	`ROUGHENOUGH_SEED`	Required	A 32-byte hexadecimal value used to generate the server's long-term key pair. This is a secret value and must be un-guessable, treat it with care. (If compiled with KMS support, length will vary)
`batch_size`	`ROUGHENOUGH_BATCH_SIZE`	Optional	The maximum number of requests to process in one batch. All nonces in a batch are used to build a Merkle tree, the root of which is signed. Default is `64` requests per batch.
`status_interval`	`ROUGHENOUGH_STATUS_INTERVAL`	Optional	Number of seconds between each logged status update. Default is `600` seconds (10 minutes).
`health_check_port`	`ROUGHENOUGH_HEALTH_CHECK_PORT`	Optional	If present, enable an HTTP health check responder on the provided port. Use with caution.
`kms_protection`	`ROUGHENOUGH_KMS_PROTECTION`	Optional	If compiled with KMS support, the ID of the KMS key used to protect the long-term identity.
`client_stats`	`ROUGHENOUGH_CLIENT_STATS`	Optional	A value of `on` or `yes` will enable tracking of per-client request statistics that will be output each time server status is logged. Default is `off` (disabled).
`fault_percentage`	`ROUGHENOUGH_FAULT_PERCENTAGE`	Optional	Likelihood (as a percentage) that the server will intentionally return an invalid client response. An integer range from `0` (disabled, all responses valid) to `50` (50% of responses will be invalid). Default is `0` (disabled).

Implementations of this trait obtain a valid configuration from different back-end sources. See:

FileConfig - configure via a YAML file
EnvironmentConfig - configure via environment variables

Required methods

`fn interface(&self) -> &str`

[Required] IP address or interface name to listen for client requests

`fn port(&self) -> u16`

[Required] UDP port to listen for requests

`fn seed(&self) -> Vec<u8>`

[Required] A 32-byte hexadecimal value used to generate the server's long-term key pair. This is a secret value and must be un-guessable, treat it with care.

`fn batch_size(&self) -> u8`

[Optional] The maximum number of requests to process in one batch. All nonces in a batch are used to build a Merkle tree, the root of which is signed. Defaults to DEFAULT_BATCH_SIZE

`fn status_interval(&self) -> Duration`

[Optional] Amount of time between each logged status update. Defaults to DEFAULT_STATUS_INTERVAL

`fn kms_protection(&self) -> &KmsProtection`

[Optional] Method used to protect the seed for the server's long-term key pair. Defaults to "plaintext" (no encryption, seed is in the clear).

`fn health_check_port(&self) -> Option<u16>`

[Optional] If present, the TCP port to respond to Google-style HTTP "legacy health check". This is a very simplistic check, it emits a fixed HTTP response to all TCP connections. https://cloud.google.com/load-balancing/docs/health-checks#legacy-health-checks

`fn client_stats_enabled(&self) -> bool`

[Optional] A value of on or yes will enable tracking of per-client request statistics that will be output each time server status is logged. Default is off (disabled).

`fn fault_percentage(&self) -> u8`

[Optional] Likelihood (as a percentage) that the server will intentionally return an invalid client response. An integer range from 0 (disabled, all responses valid) to 50 (~50% of responses will be invalid). Default is 0 (disabled).

See the Roughtime spec for background and rationale.

Loading content...