1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
//! Backends are the logical target of network traffic. They have an identity and
//! a load-balancing policy. See [Backend] to get started.
use crate::BackendId;
use serde::{Deserialize, Serialize};
#[cfg(feature = "typeinfo")]
use junction_typeinfo::TypeInfo;
/// A Backend is a logical target for network traffic.
///
/// A backend configures how all traffic for it's `target` is handled. Any
/// traffic routed to this backend will use its load balancing policy to evenly
/// spread traffic across all available endpoints.
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
#[cfg_attr(feature = "typeinfo", derive(TypeInfo))]
pub struct Backend {
/// A unique description of what this backend is.
pub id: BackendId,
/// How traffic to this target should be load balanced.
pub lb: LbPolicy,
}
// TODO: figure out how we want to support the filter_state/connection_properties style of hashing
// based on source ip or grpc channel.
//
// TODO: add support for query parameter based hashing, which involves parsing query parameters,
// which http::uri just doesn't do. switch the whole crate to url::Url or something.
//
// TODO: Random, Maglev
//
/// A policy describing how traffic to this target should be load balanced.
#[derive(Serialize, Deserialize, Clone, Debug, Default, PartialEq)]
#[serde(tag = "type")]
#[cfg_attr(feature = "typeinfo", derive(TypeInfo))]
pub enum LbPolicy {
/// A simple round robin load balancing policy. Endpoints are picked in sequential order, but
/// that order may vary client to client.
RoundRobin,
/// Use a ketama-style consistent hashing algorithm to route this request.
RingHash(RingHashParams),
/// No load balancing algorithm was specified. Clients may decide how load balancing happens
/// for this target.
#[default]
Unspecified,
}
impl LbPolicy {
/// Return `true` if this policy is [LbPolicy::Unspecified].
pub fn is_unspecified(&self) -> bool {
matches!(self, Self::Unspecified)
}
}
/// Policy for configuring a ketama-style consistent hashing algorithm.
#[derive(Serialize, Deserialize, Clone, Debug, PartialEq)]
#[serde(deny_unknown_fields)]
#[cfg_attr(feature = "typeinfo", derive(TypeInfo))]
pub struct RingHashParams {
/// The minimum size of the hash ring
#[serde(default = "default_min_ring_size", alias = "minRingSize")]
pub min_ring_size: u32,
/// How to hash an outgoing request into the ring.
///
/// Hash parameters are applied in order. If the request is missing an input, it has no effect
/// on the final hash. Hashing stops when only when all polices have been applied or a
/// `terminal` policy matches part of an incoming request.
///
/// This allows configuring a fallback-style hash, where the value of `HeaderA` gets used,
/// falling back to the value of `HeaderB`.
///
/// If no policies match, a random hash is generated for each request.
#[serde(default, skip_serializing_if = "Vec::is_empty", alias = "hashParams")]
pub hash_params: Vec<SessionAffinityHashParam>,
}
pub(crate) const fn default_min_ring_size() -> u32 {
1024
}
#[cfg(test)]
mod tests {
use serde_json::json;
use super::*;
#[test]
fn parses_lb_policy() {
let test_json = json!({
"type":"RingHash",
"minRingSize": 100
});
let obj: LbPolicy = serde_json::from_value(test_json.clone()).unwrap();
assert_eq!(
obj,
LbPolicy::RingHash(RingHashParams {
min_ring_size: 100,
hash_params: vec![]
})
);
assert_eq!(
json!({"type": "RingHash", "min_ring_size": 100}),
serde_json::to_value(obj).unwrap(),
);
}
#[test]
fn test_backend_json_roundtrip() {
let test_json = json!({
"id": { "name": "foo", "namespace": "bar", "port": 789 },
"lb": {
"type": "Unspecified",
},
});
let obj: Backend = serde_json::from_value(test_json.clone()).unwrap();
let output_json = serde_json::to_value(obj).unwrap();
assert_eq!(test_json, output_json);
}
}
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
#[serde(tag = "type")]
#[cfg_attr(feature = "typeinfo", derive(TypeInfo))]
pub enum SessionAffinityHashParamType {
/// Hash the value of a header. If the header has multiple values, they will all be used as hash
/// input.
#[serde(alias = "header")]
Header {
/// The name of the header to use as hash input.
name: String,
},
}
// FIXME: Ben votes to skip the extra "affinity" naming here as its redundant
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
#[cfg_attr(feature = "typeinfo", derive(TypeInfo))]
pub struct SessionAffinityHashParam {
/// Whether to stop immediately after hashing this value.
///
/// This is useful if you want to try to hash a value, and then fall back to another as a
/// default if it wasn't set.
#[serde(default, skip_serializing_if = "std::ops::Not::not")]
pub terminal: bool,
#[serde(flatten)]
pub matcher: SessionAffinityHashParamType,
}
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq, Default)]
#[cfg_attr(feature = "typeinfo", derive(TypeInfo))]
pub struct SessionAffinity {
#[serde(
default,
skip_serializing_if = "Vec::is_empty",
alias = "hashParams",
alias = "HashParams"
)]
pub hash_params: Vec<SessionAffinityHashParam>,
}
#[cfg(test)]
mod test_session_affinity {
use super::*;
use serde_json::json;
#[test]
fn parses_session_affinity_policy() {
let test_json = json!({
"hash_params": [
{ "type": "Header", "name": "FOO", "terminal": true },
{ "type": "Header", "name": "FOO"}
]
});
let obj: SessionAffinity = serde_json::from_value(test_json.clone()).unwrap();
let output_json = serde_json::to_value(obj).unwrap();
assert_eq!(test_json, output_json);
}
}