JMeter如何实现POST请求Payload参数化及Base64编码（对接Google Pub/Sub）

核心调整思路 #

原方案将完整明文JSON存入CSV，无法对内部字段单独做动态化（比如批量替换某字段测试值、动态生成时间戳等）。调整后不再将整段JSON作为CSV字段，而是拆分需要动态化的字段为独立CSV列，发送前临时拼接完整JSON再做Base64编码，兼顾参数灵活性和接口格式要求。

具体实现步骤 #

1. 调整CSV配置与测试数据格式 #

• 改用CSV Data Set Config拆分字段，不再存储整段JSON
• 示例CSV表头（可根据需要增减动态字段）：
  identId,requestId,requestid,timeStamp,version,eventId,creatorId,messageTTL,checkid,troubleSource,diagData
• 对应测试数据行示例：
  123456,,1,1617873956,0.0.0.1,1,0,34560,1,0,"[2020-01-01 16:00:53.707961][[lat[0]][long[0]][alt[0]][canbetrust[0]][mars[0]]][signal[5]][TEM2 wake up]"
• CSV Data Set Config需开启允许带引号数据选项，适配包含特殊字符的字段。

2. 修改前置编码脚本 #

按需选择以下两种实现方案：

场景1：保留原有BeanShell PreProcessor

直接替换原有脚本逻辑，先拼接完整JSON再编码：

import org.apache.jmeter.protocol.http.util.Base64Encoder;

// 从变量获取拆分后的参数
String identId = vars.get("identId");
String requestId = vars.get("requestId");
String requestid = vars.get("requestid");
String timeStamp = vars.get("timeStamp");
String version = vars.get("version");
String eventId = vars.get("eventId");
String creatorId = vars.get("creatorId");
String messageTTL = vars.get("messageTTL");
String checkid = vars.get("checkid");
String troubleSource = vars.get("troubleSource");
String diagData = vars.get("diagData");

// 处理空requestId为JSON原生null
String requestIdJson = (requestId != null && !requestId.isEmpty()) ? "\"" + requestId + "\"" : "null";

// 拼接完整明文JSON，注意字符串类型加双引号，数字不加，特殊字符转义
String rawPayload = "{" +
    "\"identId\":\"" + identId + "\"," +
    "\"requestId\":" + requestIdJson + "," +
    "\"payload\":{" +
        "\"header\":{" +
            "\"requestid\":" + requestid + "," +
            "\"timeStamp\":" + timeStamp + "," +
            "\"version\":\"" + version + "\"," +
            "\"eventId\":" + eventId + "," +
            "\"creatorId\":" + creatorId + "," +
            "\"messageTTL\":" + messageTTL +
        "}," +
        "\"body\":{" +
            "\"checkid\":" + checkid + "," +
            "\"checkData\":{" +
                "\"diagnosticsData\":{" +
                    "\"troubleSource\":" + troubleSource + "," +
                    "\"data\":\"" + diagData.replace("\"", "\\\"") + "\"" +
                "}" +
            "}" +
        "}" +
    "}" +
"}";

// 编码Base64并存入变量
String encodedPayload = Base64Encoder.encode(rawPayload);
vars.put("csv_payload_encoded", encodedPayload);

场景2：推荐使用JSR223 PreProcessor（Groovy）

压测场景下Groovy性能远优于BeanShell，且自带JSON序列化能力，避免手动拼接的转义错误：

import org.apache.jmeter.protocol.http.util.Base64Encoder
import groovy.json.JsonOutput

// 组装参数Map，自动处理类型适配
def params = [
    identId: vars.get("identId"),
    requestId: vars.get("requestId") ?: null,
    payload: [
        header: [
            requestid: vars.get("requestid") as int,
            timeStamp: vars.get("timeStamp") as long,
            version: vars.get("version"),
            eventId: vars.get("eventId") as int,
            creatorId: vars.get("creatorId") as int,
            messageTTL: vars.get("messageTTL") as int
        ],
        body: [
            checkid: vars.get("checkid") as int,
            checkData: [
                diagnosticsData: [
                    troubleSource: vars.get("troubleSource") as int,
                    data: vars.get("diagData")
                ]
            ]
        ]
    ]
]

// 自动生成合规JSON，无需手动转义
def rawPayload = JsonOutput.toJson(params)
def encodedPayload = Base64Encoder.encode(rawPayload)
vars.put("csv_payload_encoded", encodedPayload)

3. POST请求无需调整 #

原有Pub/Sub接口的请求体保持不变，仍使用${csv_payload_encoded}引用编码后的变量即可，符合接口格式要求。

注意事项 #

• 手动拼接JSON时需严格区分字符串、数字、null的格式，避免遗漏双引号或类型错误
• 若字段包含双引号、反斜杠等特殊字符，需手动转义，或直接使用Groovy方案规避该问题
• 压测场景优先选择JSR223 Groovy方案，减少性能损耗

内容的提问来源于stack exchange，提问作者Martin Jönsson