You need to use a generator and a provider of type grpc/scenario
pools:
- id: Pool name
gun:
type: grpc/scenario
target: localhost:8888
ammo:
type: grpc/scenario
file: payload.hcl
The minimum generator configuration is as follows
gun:
type: grpc/scenario
target: localhost:8888
For a scenario gRPC generator, all settings of a regular gRPC generator are supported gRPC generator
The provider accepts only one parameter - the path to the file with the scenario description
ammo:
type: http/scenario
file: payload.hcl
Supports file extensions
Supports formats
Several scenarios can be described in a single file. A script has a name by which one scenario differs from another.
A script is a sequence of rpc calls. That is, you will need to describe in the script which calls should be executed in what order.
The Call is a gRPC call. It has standard gRPC call fields plus additional ones. See Calls.
variable_source "users" "file/csv" {
file = "users.csv"
fields = ["user_id", "login", "pass"]
ignore_first_line = true
delimiter = ","
}
variable_source "filter_src" "file/json" {
file = "filter.json"
}
variable_source "variables" "variables" {
variables = {
header = "yandex"
b = "s"
}
}
call "auth_req" {
call = "target.TargetService.Auth"
tag = "auth"
metadata = {
"metadata" = "server.proto"
}
preprocessor "prepare" {
mapping = {
user = "source.users[next]"
}
}
payload = <<EOF
{"login": "", "pass": ""}
EOF
postprocessor "assert/response" {
payload = ["token"]
status_code = 200
}
}}
}
scenario "scenario_name" {
weight = 1
min_waiting_time = 1000
requests = [
"auth_req",
]
}
You can also see an example in the tests https://github.com/yandex/pandora/blob/dev/tests/grpc_scenario/testdata/grpc_payload.hcl
variable_sources:
- type: "file/csv"
name: "users"
ignore_first_line: true
delimiter: ","
file: "file.csv"
fields: ["user_id", "login", "pass"]
- type: "file/json"
name: "filter_src"
calls:
- name: "auth_req"
call: 'target.TargetService.Auth'
tag: auth
method: POST
metadata:
metadata: "server.proto"
preprocessors:
- type: prepare
mapping:
new_var: source.var_name[next].0
payload: '{"login": "", "pass": ""}'
postprocessors:
- type: assert/response
payload: ["token"]
status_code: 200
scenarios:
- name: scenario_name
weight: 1
min_waiting_time: 1000
requests: [
auth_req
]
Fields
The fields metadata', payload` are templated.
The standard go template is used.
Variable names have the full path of their definition.
For example
Variable users from source user_file - {{{.source.user_file.users}}
Variable item from the list_req call preprocessor - {{{.request.list_req.preprocessor.item}}
Note To maintain similarity with http scripts, the response section from the grpc call is saved to the
postprocessorsection
Variable token from the list_req call is {{{.request.list_req.postprocessor.token}}
Since the standard Go templating engine is used, it is possible to use built-in functions available at https://pkg.go.dev/text/template#hdr-Functions.
Additionally, some functions include:
For more details about randomization functions, see more.
Preprocessor - actions are performed before templating
It is used for creating new variable mapping
The preprocessor has the ability to work with arrays using modifiers
calls:
- name: req_name
...
preprocessors:
- type: prepare
mapping:
user_id: source.users[next].id
call "req_name" {
preprocessor "prepare" {
mapping = {
user_id = "source.users[next].id"
}
}
}
Checks header and body content
Upon assertion, further scenario execution is dropped
postprocessor "assert/response" {
payload = ["token"]
status_code = 200
}
This section repeats the same scenario in HTTP generator
The minimum fields for the script are name and list of requests
scenario "scenario_name" {
requests = [
"list_call",
"order_call",
"order_call",
"order_call"
]
}
More - scenario in HTTP generator
Follow - Variable sources