Abp vNext異常處理的缺陷/改造方案

之前吐槽Abp的用戶/租戶管理模塊！今天我又來了，這次我給Abp官方repo提了一個issue。

目前Website使用Abp vNext開發(fā)，免不了要全局處理異常、提示服務器異常信息。

1. Abp官方異常處理

Abp項目默認會啟動內置的異常處理，默認不將異常信息發(fā)送到客戶端。
在AppModule文件ConfigureServices方法中使用以下代碼：

Configure(options?=>
{
????options.SendExceptionsDetailsToClients?=?true;
});

可將異常信息發(fā)送到客戶端：如下圖：

{
"error":?{
   "code":?null,
   "message":?"ERROR?[42000]?[Cloudera][ImpalaODBC]?(360)?Syntax?error?occurred?during?query?execution:?[HY000]?:?AnalysisException:?Could?not?resolve?column/field?reference:?'ug_fed89221846a42dc8427932b2965a020'\n",
   "details":?"OdbcException:?ERROR?[42000]?[Cloudera][ImpalaODBC]?(360)?Syntax?error?occurred?during?query?execution:?[HY000]?:?AnalysisException:?Could?not?resolve?column/field?reference:?'ug_fed89221846a42dc8427932b2965a020'\n\nSTACK?TRACE:?at?Gridsum.EAP.Olap.ExecuteQueryLayer.HandleQueryAsync(QueryContext?queryContext,?CancellationToken?cancellationToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/Olap/ExecuteQueryLayer.cs:line?81\n?at?Gridsum.EAP.Olap.DistributedCacheLayer.HandleQueryAsync(QueryContext?queryContext,?CancellationToken?cancellationToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/Olap/DistributedCacheLayer.cs:line?50\n?at?Gridsum.EAP.DataQuery.AbstractQueryExecutor`1.ExecuteQueryAsync(TQuery?query,?DistributedCacheEntryOptions?options,?CancellationToken?cancellationToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/AbstractQueryExecutor.cs:line?60\n?at?Gridsum.EAP.DataQuery.AbstractQueryExecutor`1.ExecuteQueryAsync(TQuery?query,?CancellationToken?token)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/AbstractQueryExecutor.cs:line?47\n?at?Gridsum.EAP.Application.UserGroupService.ClearUserGroupUserAsync(UserGroupUpdateUserDto?updateDto,?String?idshort,?CancellationToken?cancelToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.Application/UserGroup/UserGroupService.cs:line?332\n?at?Gridsum.EAP.Application.UserGroupService.UpdateUserGroupUserAsync(UserGroupUpdateUserDto?updateDto,?CancellationToken?cancelToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.Application/UserGroup/UserGroupService.cs:line?370\n?at?Castle.DynamicProxy.AsyncInterceptorBase.ProceedAsynchronous[TResult](IInvocation?invocation,?IInvocationProceedInfo?proceedInfo?"42000]?[Cloudera][ImpalaODBC]?(360?"42000]?[Cloudera][ImpalaODBC]?(360)?Syntax?error?occurred?during?query?execution:?[HY000]?:?AnalysisException:?Could?not?resolve?column/field?reference:?'ug_fed89221846a42dc8427932b2965a020'\n\nSTACK?TRACE:?at?Gridsum.EAP.Olap.ExecuteQueryLayer.HandleQueryAsync(QueryContext?queryContext,?CancellationToken?cancellationToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/Olap/ExecuteQueryLayer.cs:line?81\n?at?Gridsum.EAP.Olap.DistributedCacheLayer.HandleQueryAsync(QueryContext?queryContext,?CancellationToken?cancellationToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/Olap/DistributedCacheLayer.cs:line?50\n?at?Gridsum.EAP.DataQuery.AbstractQueryExecutor`1.ExecuteQueryAsync(TQuery?query,?DistributedCacheEntryOptions?options,?CancellationToken?cancellationToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/AbstractQueryExecutor.cs:line?60\n?at?Gridsum.EAP.DataQuery.AbstractQueryExecutor`1.ExecuteQueryAsync(TQuery?query,?CancellationToken?token)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/AbstractQueryExecutor.cs:line?47\n?at?Gridsum.EAP.Application.UserGroupService.ClearUserGroupUserAsync(UserGroupUpdateUserDto?updateDto,?String?idshort,?CancellationToken?cancelToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.Application/UserGroup/UserGroupService.cs:line?332\n?at?Gridsum.EAP.Application.UserGroupService.UpdateUserGroupUserAsync(UserGroupUpdateUserDto?updateDto,?CancellationToken?cancelToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.Application/UserGroup/UserGroupService.cs:line?370\n?at?Castle.DynamicProxy.AsyncInterceptorBase.ProceedAsynchronous[TResult")?Syntax?error?occurred?during?query?execution:?[HY000]?:?AnalysisException:?Could?not?resolve?column/field?reference:?'ug_fed89221846a42dc8427932b2965a020'\n\nSTACK?TRACE:?at?Gridsum.EAP.Olap.ExecuteQueryLayer.HandleQueryAsync(QueryContext?queryContext,?CancellationToken?cancellationToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/Olap/ExecuteQueryLayer.cs:line?81\n?at?Gridsum.EAP.Olap.DistributedCacheLayer.HandleQueryAsync(QueryContext?queryContext,?CancellationToken?cancellationToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/Olap/DistributedCacheLayer.cs:line?50\n?at?Gridsum.EAP.DataQuery.AbstractQueryExecutor`1.ExecuteQueryAsync(TQuery?query,?DistributedCacheEntryOptions?options,?CancellationToken?cancellationToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/AbstractQueryExecutor.cs:line?60\n?at?Gridsum.EAP.DataQuery.AbstractQueryExecutor`1.ExecuteQueryAsync(TQuery?query,?CancellationToken?token)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.DataQuery/AbstractQueryExecutor.cs:line?47\n?at?Gridsum.EAP.Application.UserGroupService.ClearUserGroupUserAsync(UserGroupUpdateUserDto?updateDto,?String?idshort,?CancellationToken?cancelToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.Application/UserGroup/UserGroupService.cs:line?332\n?at?Gridsum.EAP.Application.UserGroupService.UpdateUserGroupUserAsync(UserGroupUpdateUserDto?updateDto,?CancellationToken?cancelToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.Application/UserGroup/UserGroupService.cs:line?370\n?at?Castle.DynamicProxy.AsyncInterceptorBase.ProceedAsynchronous[TResult")\n?at?Volo.Abp.Castle.DynamicProxy.CastleAbpMethodInvocationAdapterWithReturnValue`1.ProceedAsync()\n?at?Volo.Abp.Validation.ValidationInterceptor.InterceptAsync(IAbpMethodInvocation?invocation)\n?at?Volo.Abp.Castle.DynamicProxy.CastleAsyncAbpInterceptorAdapter`1.InterceptAsync[TResult](IInvocation?invocation,?IInvocationProceedInfo?proceedInfo,?Func`3?proceed?"TResult")\n?at?Castle.DynamicProxy.AsyncInterceptorBase.ProceedAsynchronous[TResult](IInvocation?invocation,?IInvocationProceedInfo?proceedInfo?"TResult")\n?at?Volo.Abp.Castle.DynamicProxy.CastleAbpMethodInvocationAdapterWithReturnValue`1.ProceedAsync()\n?at?Volo.Abp.Auditing.AuditingInterceptor.InterceptAsync(IAbpMethodInvocation?invocation)\n?at?Volo.Abp.Castle.DynamicProxy.CastleAsyncAbpInterceptorAdapter`1.InterceptAsync[TResult](IInvocation?invocation,?IInvocationProceedInfo?proceedInfo,?Func`3?proceed?"TResult")\n?at?Castle.DynamicProxy.AsyncInterceptorBase.ProceedAsynchronous[TResult](IInvocation?invocation,?IInvocationProceedInfo?proceedInfo?"TResult")\n?at?Volo.Abp.Castle.DynamicProxy.CastleAbpMethodInvocationAdapterWithReturnValue`1.ProceedAsync()\n?at?Volo.Abp.Uow.UnitOfWorkInterceptor.InterceptAsync(IAbpMethodInvocation?invocation)\n?at?Volo.Abp.Castle.DynamicProxy.CastleAsyncAbpInterceptorAdapter`1.InterceptAsync[TResult](IInvocation?invocation,?IInvocationProceedInfo?proceedInfo,?Func`3?proceed?"TResult")\n?at?Gridsum.EAP.Controllers.UserGroupController.UpdateUserGroupUserAsync(String?id,?CancellationToken?cancelToken)?in?/home/gitlab-runner/builds/ttRjAPVA/0/eap/website/app/src/Gridsum.EAP.HttpApi/Controllers/UserGroupController.cs:line?320\n?at?lambda_method3440(Closure?,?Object?)\n?at?Microsoft.AspNetCore.Mvc.Infrastructure.ActionMethodExecutor.AwaitableObjectResultExecutor.Execute(IActionResultTypeMapper?mapper,?ObjectMethodExecutor?executor,?Object?controller,?Object[]?arguments)\n?at?Microsoft.AspNetCore.Mvc.Infrastructure.ControllerActionInvoker.g__Awaited|12_0(ControllerActionInvoker?invoker,?ValueTask`1?actionResultValueTask)\n?at?Microsoft.AspNetCore.Mvc.Infrastructure.ControllerActionInvoker.g__Awaited|10_0(ControllerActionInvoker?invoker,?Task?lastTask,?State?next,?Scope?scope,?Object?state,?Boolean?isCompleted)\n?at?Microsoft.AspNetCore.Mvc.Infrastructure.ControllerActionInvoker.Rethrow(ActionExecutedContextSealed?context)\n?at?Microsoft.AspNetCore.Mvc.Infrastructure.ControllerActionInvoker.Next(State&?next,?Scope&?scope,?Object&?state,?Boolean&?isCompleted)\n?at?Microsoft.AspNetCore.Mvc.Infrastructure.ControllerActionInvoker.g__Awaited|13_0(ControllerActionInvoker?invoker,?Task?lastTask,?State?next,?Scope?scope,?Object?state,?Boolean?isCompleted)\n?at?Microsoft.AspNetCore.Mvc.Infrastructure.ResourceInvoker.g__Awaited|25_0(ResourceInvoker?invoker,?Task?lastTask,?State?next,?Scope?scope,?Object?state,?Boolean?isCompleted)\n",
   "data":?null,
   "validationErrors":?null
  }
}

經過幾天倒騰，發(fā)現(xiàn)Abp vNext的異常處理有幾個問題。

2.Abp異常處理存在的缺陷

1. 并沒有如官方所述：自動處理所有異常，實際需要滿足官方所說的某個條件：

這就導致當Controller Action方法返回的不是object result時，則根本捕獲不到異常(我們暫時不說middleware產生的異常)，這應該算Abp的一個Bug

2. 輸出的異常沒有TraceId， 不利于日志排查
3. 發(fā)送到客戶端的日志字段message,detail過于詳細冗長，不適合前端顯示

也可以配置SendExceptionsDetailsToClients，不將異常信息發(fā)送到客戶端，但這樣就因噎廢食了。

3. 異常處理的目標

雖然Abp的異常處理有缺陷， 但只是異常信息應用上的缺陷，
Abp異常處理①對異常的劃分、②異常信息的本地化、③出現(xiàn)異常時寫日志 ?支持的還是相當好。

基于Abp的異常處理現(xiàn)狀，考慮做一些改進：

1. 對所有Controller-Action方法捕獲異常， [修復Abp Bug]
2. 在Abp的異常處理結果中添加 TraceId
3. 希望將服務端異常分類，簡化后給到前端；同時也不妨礙開發(fā)者查看詳細異常信息。

4. 揪出Abp異常處理缺陷的根源

Abp異常處理的核心對象AbpExceptionFilter，實現(xiàn)IAsyncExceptionFilter過濾器, ITransientDependency瞬時注入接口。

這是一個ServiceFilterAttribute, 你可以理解有個特性作用在每一個Controller的Action方法上：

[ServiceFilter(typeof(AbpExceptionFilter))]

一旦某個Controller的Action方法發(fā)生異常， 會執(zhí)行如下代碼：

public?async?Task?OnExceptionAsync(ExceptionContext?context)
{
???if?(!ShouldHandleException(context))
???{
????????return;
???}
???await?HandleAndWrapException(context);
}

1. ShouldHandleException(context) ：監(jiān)測是否應該處理異常，據查該函數(shù)確實存在我上文說的問題：并不能捕獲所有的Action方法的異常。
2. HandleAndWrapException(context)：異常處理步驟：
• 根據Abp內置的異常類型，自動確定狀態(tài)碼 (這個在Abp官方文檔有講)
• 序列化異常對象，并向客戶端輸出如下格式：

{
"error":?{
  "code":?null,
   "message":?"ERROR?[42000]?[Cloudera][ImpalaODBC]?(360)?Syntax?error?occurred?during?query?execution:?[HY000]?:?AnalysisException:?Could?not?resolve?column/field?reference:?'ug_fed89221846a42dc8427932b2965a020'\n",
   "details":xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx......,
   "data":?null,
   "validationErrors":?null
  }
}

① 輸出的信息，從一開始就沒有包含TraceId；
② Abp標準格式化后的異常信息，過于冗長，message，details字段均不適合前端顯示。

• 寫日志，默認異常級別為Error

掌握以上源碼，我們可以針對性的改造Abp的核心異常處理類AbpExceptionFilter。

5. Abp異常處理: 缺陷修復方案

光說不練假把式

Abp的AbpExceptionFilter不是抽象類，沒法重載，為達到我們設定的3個目標。
考慮使用針對性的ExceptionFilter替換默認有缺陷的AbpExceptionFilter。

①.? 新建EapExceptionFilter，內容拷貝自AbpExceptionFilter， 并做出如下針對性修改：

②. 在AppModule中，替換默認的AbpExceptionFilter為新的EapExceptionFilter過濾器：

context.Services.AddMvc(options?=>
{
????options.Filters.ReplaceOne(x=>?(x?as?ServiceFilterAttribute)?.ServiceType?.Name==nameof(AbpExceptionFilter),?new?ServiceFilterAttribute(typeof(EapExceptionFilter)));??
})

改造的效果如下：

That's? All

如果大家真切使用了Abp vNext最新版，

相信我在第2點提到的Abp異常處理的缺陷，Abp使用者會感同身受；

第3點提出的幾個目標也是企業(yè)級異常處理要解決的痛點。

此異常處理的思路也可推及到其他非Abp項目.

改造方案在Abp官方github issue上：? https://github.com/abpframework/abp/issues/6761

一家之言，如有其他看法，請不吝賜教！

(btw，公眾號文章發(fā)布之后，限制修改；若有后續(xù)，請 [閱讀原文]！)

Reference

1. https://github.com/abpframework/abp/blob/dev/framework/src/Volo.Abp.AspNetCore.Mvc/Volo/Abp/AspNetCore/Mvc/ExceptionHandling/AbpExceptionFilter.cs
2. https://docs.abp.io/zh-Hans/abp/latest/Exception-Handling
   
3. https://docs.microsoft.com/en-us/aspnet/core/mvc/controllers/filters?view=aspnetcore-5.0
   
   
   閱讀更多