Struct agreed::config::Config

pub struct Config {
    pub cluster_name: String,
    pub election_timeout_min: u64,
    pub election_timeout_max: u64,
    pub heartbeat_interval: u64,
    pub max_payload_entries: u64,
    pub replication_lag_threshold: u64,
    pub snapshot_policy: SnapshotPolicy,
    pub snapshot_max_chunk_size: u64,
    pub catch_up_cancellation_policy: CatchUpCancellationPolicy,
}

The runtime configuration for a Raft node.

The default values used by this type should generally work well for Raft clusters which will be running with nodes in multiple datacenter availability zones with low latency between zones. These values should typically be made configurable from the perspective of the application which is being built on top of Raft.

When building the Raft configuration for your application, remember this inequality from the Raft spec: broadcastTime ≪ electionTimeout ≪ MTBF.

In this inequality broadcastTime is the average time it takes a server to send RPCs in parallel to every server in the cluster and receive their responses; electionTimeout is the election timeout described in Section 5.2; and MTBF is the average time between failures for a single server. The broadcast time should be an order of magnitude less than the election timeout so that leaders can reliably send the heartbeat messages required to keep followers from starting elections; given the randomized approach used for election timeouts, this inequality also makes split votes unlikely. The election timeout should be a few orders of magnitude less than MTBF so that the system makes steady progress. When the leader crashes, the system will be unavailable for roughly the election timeout; we would like this to represent only a small fraction of overall time.

What does all of this mean? Simply keep your election timeout settings high enough that the performance of your network will not cause election timeouts, but don’t keep it so high that a real leader crash would cause prolonged downtime. See the Raft spec §5.6 for more details.

Fields

cluster_name: String

The application specific name of this Raft cluster.

This does not influence the Raft protocol in any way, but is useful for observability.

election_timeout_min: u64

The minimum election timeout in milliseconds.

election_timeout_max: u64

The maximum election timeout in milliseconds.

heartbeat_interval: u64

The heartbeat interval in milliseconds at which leaders will send heartbeats to followers.

Defaults to 50 milliseconds.

NOTE WELL: it is very important that this value be greater than the amount if time it will take on average for heartbeat frames to be sent between nodes. No data processing is performed for heartbeats, so the main item of concern here is network latency. This value is also used as the default timeout for sending heartbeats.

max_payload_entries: u64

The maximum number of entries per payload allowed to be transmitted during replication.

When configuring this value, it is important to note that setting this value too low could cause sub-optimal performance. This will primarily impact the speed at which slow nodes, nodes which have been offline, or nodes which are new to the cluster, are brought up-to-speed. If this is too low, it will take longer for the nodes to be brought up to consistency with the rest of the cluster.

replication_lag_threshold: u64

The distance behind in log replication a follower must fall before it is considered “lagging”.

This configuration parameter controls replication streams from the leader to followers in the cluster. Once a replication stream is considered lagging, it will stop buffering entries being replicated, and instead will fetch entries directly from the log until it is up-to-speed, at which time it will transition out of “lagging” state back into “line-rate” state.

snapshot_policy: SnapshotPolicy

The snapshot policy to use for a Raft node.

snapshot_max_chunk_size: u64

The maximum snapshot chunk size allowed when transmitting snapshots (in bytes).

Defaults to 3Mib.

catch_up_cancellation_policy: CatchUpCancellationPolicy

The cancellation policy of the configuration change catch-up process used on this Raft node.

Implementations

impl Config

pub fn build(cluster_name: String) -> ConfigBuilder

Start the builder process for a new Config instance. Call validate when done.

The directory where the log snapshots are to be kept for a Raft node is required and must be specified to start the config builder process.

pub fn new_rand_election_timeout(&self) -> u64

Generate a new random election timeout within the configured min & max.

Trait Implementations

impl Debug for Config

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Config

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error> where
    __D: Deserializer<'de>, 

Deserialize this value from the given Serde deserializer.

impl Serialize for Config

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error> where
    __S: Serializer, 

Serialize this value into the given Serde serializer.