在上一篇博客 《Zipkin原理学习–日志追踪 MySQL 执行语句》 中我们已经了解学习到 Zipkin 官方提供的针对 MySQL 数据库 sql 语句执行的追踪拦截器，现在我们基于数据库连接池 Druid 的 Filter 机制 写一个能支持多种数据库（mysql，pg、oracle 等）日志追踪拦截器。
Druid 过滤器 Filter
DruidDataSource支持通过Filter-Chain模式进行扩展，类似Serlvet的Filter，扩展十分方便，你可以拦截任何JDBC的方法。
有两种配置Filter的方式，一种是配置filters属性，一种是配置proxyFilters属性。filters和proxyFilters的配置是组合关系，而不是替换关系。
配置filters属性
配置filters属性比较简单，filters的类型是字符串，多个filter使用逗号隔开。例如：
<bean id="dataSource" class="com.alibaba.druid.pool.DruidDataSource"
init-method="init" destroy-method="close">
<property name="url" value="jdbc:derby:memory:spring-test;create=true" />
<property name="initialSize" value="1" />
<property name="maxActive" value="20" />
<property name="filters" value="stat,log4j" />
</bean>
filters属性的配置使用别名或者全类名，stat是com.alibaba.druid.filter.stat.StatFilter的别名。 内置Filter的别名 查看内置Filter的别名。
配置proxyFilters属性
proxyFilters的类型是List，使用proxyFilters配置，可以有更多的配置选项。
<bean id="stat-filter" class="com.alibaba.druid.filter.stat.StatFilter">
</bean>
<bean id="dataSource" class="com.alibaba.druid.pool.DruidDataSource"
init-method="init" destroy-method="close">
<property name="url" value="jdbc:derby:memory:spring-test;create=true" />
<property name="initialSize" value="1" />
<property name="maxActive" value="20" />
<property name="proxyFilters">
<list>
<ref bean="stat-filter" />
</list>
</property>
</bean>
Zipkin 日志追踪过滤器 TracingStatementFilter
TracingStatementFilter 继承抽象类 FilterEventAdapter 实现以下接口：
protected void statementExecuteUpdateBefore(StatementProxy statement, String sql) {
}
protected void statementExecuteUpdateAfter(StatementProxy statement, String sql, int updateCount) {
}
protected void statementExecuteQueryBefore(StatementProxy statement, String sql) {
}
protected void statementExecuteQueryAfter(StatementProxy statement, String sql, ResultSetProxy resultSet) {
}
protected void statementExecuteBefore(StatementProxy statement, String sql) {
}
protected void statementExecuteAfter(StatementProxy statement, String sql, boolean result) {
}
protected void statementExecuteBatchBefore(StatementProxy statement) {
}
protected void statementExecuteBatchAfter(StatementProxy statement, int[] result) {
}
protected void statement_executeErrorAfter(StatementProxy statement, String sql, Throwable error) {
}
实现以上接口，在执行SQL语句之前、之后或抛出异常时都会执行上面接口方法，这样就可以追踪 SQL 语句的执行过程了。
完整 Zipkin 日志追踪代码：
public class TracingStatementFilter extends FilterEventAdapter {
private String zipkinServiceName;
public TracingStatementFilter(String zipkinServiceName){
this.zipkinServiceName = zipkinServiceName;
}
public TracingStatementFilter(){
}
@Override
protected void statementExecuteUpdateBefore(StatementProxy statement, String sql) {
super.statementExecuteUpdateBefore(statement,sql);
try {
Before(statement,sql);
}catch (Exception e){
}
}
@Override
protected void statementExecuteUpdateAfter(StatementProxy statement, String sql, int updateCount) {
super.statementExecuteUpdateAfter(statement,sql,updateCount);
try {
After(statement,sql);
}catch (Exception e){
}
}
@Override
protected void statementExecuteQueryBefore(StatementProxy statement, String sql) {
super.statementExecuteQueryBefore(statement,sql);
try {
Before(statement,sql);
}catch (Exception e){
}
}
@Override
protected void statementExecuteQueryAfter(StatementProxy statement, String sql, ResultSetProxy resultSet) {
super.statementExecuteQueryAfter(statement,sql,resultSet);
try {
After(statement,sql);
}catch (Exception e){
}
}
@Override
protected void statementExecuteBefore(StatementProxy statement, String sql) {
super.statementExecuteBefore(statement,sql);
try {
Before(statement,sql);
}catch (Exception e){
}
}
@Override
protected void statementExecuteAfter(StatementProxy statement, String sql, boolean result) {
super.statementExecuteAfter(statement,sql,result);
try {
After(statement,sql);
}catch (Exception e){
}
}
@Override
protected void statementExecuteBatchBefore(StatementProxy statement) {
super.statementExecuteBatchBefore(statement);
try {
Before(statement , null);
}catch (Exception e){
}
}
@Override
protected void statementExecuteBatchAfter(StatementProxy statement, int[] result) {
super.statementExecuteBatchAfter(statement , result);
try {
After(statement , null);
}catch (Exception e){
}
}
@Override
protected void statement_executeErrorAfter(StatementProxy statement, String sql, Throwable error) {
super.statement_executeErrorAfter(statement , sql , error);
try {
ErrorAfter(statement , sql , error);
}catch (Exception e){
}
}
protected void Before(StatementProxy statement, String sql) {
Span span = ThreadLocalSpan.CURRENT_TRACER.next();
if (span == null || span.isNoop()) {
return;
}
if(sql == null){
sql = statement.getLastExecuteSql();
}
// Allow span names of single-word statements like COMMIT
int spaceIndex = sql.indexOf(' ');
span.kind(Span.Kind.CLIENT).name(spaceIndex == -1 ? sql : sql.substring(0, spaceIndex));
span.tag("sql.query", sql);
parseServerIpAndPort(statement,span);
span.start();
}
protected void After(StatementProxy statement, String sql) {
Span span = ThreadLocalSpan.CURRENT_TRACER.remove();
if (span == null || span.isNoop()) {
return;
}
span.finish();
return ;
}
protected void ErrorAfter(StatementProxy statement, String sql, Throwable error) {
Span span = ThreadLocalSpan.CURRENT_TRACER.remove();
if (span == null || span.isNoop()) {
return;
}
if (error instanceof SQLException) {
span.tag("error", Integer.toString(((SQLException) error).getErrorCode()));
}
span.finish();
return ;
}
public void parseServerIpAndPort(StatementProxy statement, Span span) {
try {
URI url = URI.create(statement.getConnection().getMetaData().getURL().substring(5));
if (getZipkinServiceName() == null || "".equals(getZipkinServiceName())) {
try {
zipkinServiceName = "DB"+url.getPath();
}catch (Exception e){
;
}
}
span.remoteServiceName(getZipkinServiceName());
String host = url.getHost();
if (host != null) {
span.remoteIpAndPort(host, url.getPort());
}
} catch (Exception e) {
// remote address is optional
}
}
public String getZipkinServiceName() {
return zipkinServiceName;
}
public void setZipkinServiceName(String zipkinServiceName) {
this.zipkinServiceName = zipkinServiceName;
}
}
总结：
简单来说就是利用 Druid 提供的过滤器机制，在 SQL 语句执行前后添加追踪日志。