traffic-split
Note: The ratio between each upstream may not so accurate since the drawback of weighted round robin algorithm (especially when the wrr state is reset).
Currently, in the configuration of weighted_upstreams.upstream
, the unsupported fields are: service_name, discovery_type, checks, retries, retry_timeout, desc, scheme, labels, create_time and update_time. But you can use weighted_upstreams.upstream_id
to bind the upstream
object to achieve their functions.
The traffic-split plugin is mainly composed of two parts: match
and weighted_upstreams
. match
is a custom conditional rule, and weighted_upstreams
is upstream configuration information. If you configure match
and weighted_upstreams
information, then after the match
rule is verified, it will be based on the weight
value in weighted_upstreams
; the ratio of traffic between each upstream in the plugin will be guided, otherwise, all traffic will be directly Reach the upstream
configured on route
or service
. Of course, you can also configure only the weighted_upstreams
part, which will directly guide the traffic ratio between each upstream in the plugin based on the weight
value in weighted_upstreams
.
Note: 1. In match
, the expression in vars is the relationship of and
, and the relationship between multiple vars
is the relationship of or
. 2. In the weighted_upstreams field of the plugin, if there is a structure with only weight
, it means the upstream traffic weight value on route
or service
. Such as:
Create a route and enable the traffic-split
plugin. When configuring the upstream information of the plugin, there are two ways:
- Configure upstream information through the
upstream
attribute in the plugin.
curl http://127.0.0.1:9080/apisix/admin/routes/1 -H 'X-API-KEY: edd1c9f034335f136f87ad84b625c8f1' -X PUT -d '
{
"uri": "/index.html",
"plugins": {
"traffic-split": {
"rules": [
{
"weighted_upstreams": [
{
"upstream": {
"name": "upstream_A",
"type": "roundrobin",
"nodes": {
"127.0.0.1:1981":10
},
"timeout": {
"connect": 15,
"send": 15,
"read": 15
}
},
"weight": 1
},
{
"weight": 1
}
]
}
]
}
},
"upstream": {
"type": "roundrobin",
"nodes": {
"127.0.0.1:1980": 1
}
}
}'
- Use the
upstream_id
attribute in the plugin to bind upstream.
curl http://127.0.0.1:9080/apisix/admin/routes/1 -H 'X-API-KEY: edd1c9f034335f136f87ad84b625c8f1' -X PUT -d '
{
"uri": "/index.html",
"plugins": {
"traffic-split": {
"rules": [
{
"weighted_upstreams": [
{
"upstream_id": 1,
"weight": 1
},
{
"weight": 1
}
]
}
]
}
},
"upstream": {
"type": "roundrobin",
"nodes": {
"127.0.0.1:1980": 1
}
}
}'
The match
rule part is missing, and the traffic is split according to the weight
value configured by the weighted_upstreams
in the plugin. Divide plugin's upstream
and route's upstream
according to the traffic ratio of 3:2, of which 60% of the traffic reaches the upstream of the 1981
port in the plugin, and 40% of the traffic reaches the default 1980
port on the route Upstream.
curl http://127.0.0.1:9080/apisix/admin/routes/1 -H 'X-API-KEY: edd1c9f034335f136f87ad84b625c8f1' -X PUT -d '
{
"uri": "/index.html",
"plugins": {
"traffic-split": {
"rules": [
{
"weighted_upstreams": [
{
"upstream": {
"name": "upstream_A",
"type": "roundrobin",
"nodes": {
"127.0.0.1:1981":10
},
"timeout": {
"connect": 15,
"send": 15,
"read": 15
}
},
"weight": 3
},
{
"weight": 2
}
]
}
]
}
},
"upstream": {
"type": "roundrobin",
"nodes": {
}
}
}'
Test plugin:
There are 5 requests, 3 requests hit the upstream of port 1981 of the plugin, and 2 requests hit the upstream of port 1980 of route
.
$ curl http://127.0.0.1:9080/index.html -i
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
hello 1980
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
world 1981
......
Get the match
rule parameter through the request header (you can also get it through the request parameter or NGINX variable). After the match
rule is matched, it means that all requests hit the upstream configured by the plugin, otherwise the request only hits the route
configured upstream.
curl http://127.0.0.1:9080/apisix/admin/routes/1 -H 'X-API-KEY: edd1c9f034335f136f87ad84b625c8f1' -X PUT -d '
{
"uri": "/index.html",
"plugins": {
"traffic-split": {
"rules": [
{
"match": [
{
"vars": [
["http_release","==","new_release"]
]
}
],
"weighted_upstreams": [
{
"upstream": {
"name": "upstream_A",
"type": "roundrobin",
"nodes": {
"127.0.0.1:1981":10
}
}
}
]
}
]
}
},
"upstream": {
"type": "roundrobin",
"nodes": {
"127.0.0.1:1980": 1
}
}
}'
The rule of match
is matched, and all requests hit the upstream port 1981 configured by the plugin:
$ curl http://127.0.0.1:9080/index.html -H 'release: new_release' -i
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
......
world 1981
The match
rule fails to match, and all requests hit the 1980 port upstream configured on the route
:
Multiple vars
rules can be set in match
. Multiple expressions in vars
have an and
relationship, and multiple vars
rules have an or
relationship; as long as one of the vars is required If the rule passes, the entire match
passes.
Example 1: Only one vars
rule is configured, and multiple expressions in vars
are in the relationship of and
. In weighted_upstreams
, the traffic is divided into 3:2 according to the value of weight
, of which only the part of the weight
value represents the proportion of upstream on the route
. When match
fails to pass, all traffic will only hit the upstream on the route.
curl http://127.0.0.1:9080/apisix/admin/routes/1 -H 'X-API-KEY: edd1c9f034335f136f87ad84b625c8f1' -X PUT -d '
{
"uri": "/index.html",
"plugins": {
"traffic-split": {
"rules": [
{
"match": [
{
"vars": [
["arg_name","==","jack"],
["http_user-id",">","23"],
["http_apisix-key","~~","[a-z]+"]
]
}
],
"weighted_upstreams": [
{
"upstream": {
"name": "upstream_A",
"type": "roundrobin",
"nodes": {
"127.0.0.1:1981":10
}
},
"weight": 3
},
{
"weight": 2
}
]
}
]
}
},
"upstream": {
"type": "roundrobin",
"nodes": {
"127.0.0.1:1980": 1
}
}
}'
The plugin sets the requested match
rule and upstream with port 1981
, and the route has upstream with port 1980
.
Test plugin:
The match rule is successfully verified, and the upstream port of 1981
is hit.
$ curl 'http://127.0.0.1:9080/index.html?name=jack' -H 'user-id:30' -H 'apisix-key: hello' -i
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
......
world 1981
The match rule fails to verify, and it hits the upstream of the default port of 1980
.
$ curl 'http://127.0.0.1:9080/index.html?name=jack' -H 'user-id:30' -H 'apisix-key: hello' -i
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
......
hello 1980
After 5 requests, the service of port 1981
was hit 3 times, and the service of port 1980
was hit 2 times.
Example 2: Configure multiple vars
rules. Multiple expressions in vars
are and
relationships, and multiple vars
are or
relationships. According to the weight
value in weighted_upstreams
, the traffic is divided into 3:2, where only the part of the weight
value represents the proportion of upstream on the route. When match
fails to pass, all traffic will only hit the upstream on the route.
curl http://127.0.0.1:9080/apisix/admin/routes/1 -H 'X-API-KEY: edd1c9f034335f136f87ad84b625c8f1' -X PUT -d '
{
"uri": "/index.html",
"plugins": {
"traffic-split": {
{
"match": [
{
"vars": [
["arg_name","==","jack"],
["http_user-id",">","23"],
["http_apisix-key","~~","[a-z]+"]
]
},
{
"vars": [
["http_user-id2","!",">","33"],
["http_apisix-key2","~~","[a-z]+"]
]
}
],
"weighted_upstreams": [
{
"upstream": {
"name": "upstream_A",
"type": "roundrobin",
"nodes": {
"127.0.0.1:1981":10
}
},
"weight": 3
},
{
"weight": 2
}
]
}
]
}
},
"upstream": {
"type": "roundrobin",
"nodes": {
"127.0.0.1:1980": 1
}
}
}'
Test plugin:
$ curl 'http://127.0.0.1:9080/index.html?name=jack&name2=rose' -H 'user-id:30' -H 'user-id2:22' -H 'apisix-key: hello' -H 'apisix-key2: world' -i
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
......
world 1981
$ curl 'http://127.0.0.1:9080/index.html?name=jack&name2=rose' -H 'user-id:30' -H 'user-id2:22' -H 'apisix-key: hello' -H 'apisix-key2: world' -i
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
......
hello 1980
After 5 requests, the service of port 1981
was hit 3 times, and the service of port 1980
was hit 2 times.
$ curl 'http://127.0.0.1:9080/index.html?name=jack' -H 'user-id:30' -H 'user-id2:22' -H 'apisix-key: hello' -H 'apisix-key2: world' -i
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
......
hello 1980
After 5 requests, the service of port 1981
was hit 3 times, and the service of port 1980
was hit 2 times.
$ curl 'http://127.0.0.1:9080/index.html?name=jack' -i
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
......
hello 1980
By configuring multiple rules
, we can achieve one-to-one correspondence between different matching rules and upstream.
Example:
When the request header x-api-id
is equal to 1, it hits the upstream with port 1981; when x-api-id
is equal to 2, it hits the upstream with port 1982; otherwise, it hits the upstream with port 1980 (the upstream response data is the corresponding port number).
curl http://127.0.0.1:9080/apisix/admin/routes/1 -H 'X-API-KEY: edd1c9f034335f136f87ad84b625c8f1' -X PUT -d '
{
"uri": "/hello",
"plugins": {
"traffic-split": {
"rules": [
{
"match": [
{
"vars": [
["http_x-api-id","==","1"]
]
}
],
"weighted_upstreams": [
{
"upstream": {
"name": "upstream-A",
"type": "roundrobin",
"nodes": {
"127.0.0.1:1981":1
}
},
"weight": 3
}
]
},
{
"match": [
{
"vars": [
["http_x-api-id","==","2"]
]
}
],
"weighted_upstreams": [
{
"upstream": {
"name": "upstream-B",
"type": "roundrobin",
"nodes": {
"127.0.0.1:1982":1
}
},
"weight": 3
}
]
}
]
}
},
"upstream": {
"type": "roundrobin",
"nodes": {
"127.0.0.1:1980": 1
}
}
}'
Test plugin:
The request header x-api-id
is equal to 1, hitting the upstream with the 1981 port.
$ curl http://127.0.0.1:9080/hello -H 'x-api-id: 1'
1981
The request header x-api-id
is equal to 2, hitting the upstream with the 1982 port.
$ curl http://127.0.0.1:9080/hello -H 'x-api-id: 2'
1982
The request header x-api-id
is equal to 3, the rule does not match, and it hits the upstream with port 1980.
1980