这是Application Insight/Azure功能错误吗?或我的理解不正确
今天,我将Azure功能与应用程序见解集成在一起,以进行应用程序日志记录,尤其是捕获错误堆栈跟踪.
Today I integrate my Azure function with Application insight for application logging, especially to capture error stack-trace.
首先,我编写了没有try-catch块的Azure函数,因此它在监视器"部分和应用程序洞察中也显示了正确的状态/信息.
First I have written Azure function without try-catch block so It was showing correct status/Information in Monitor section and application insight as well.
稍后,我添加了try-catch块并记录了更多数据,例如
Later I added try-catch block and logged some more data like
catch(Exception ex)
{
log.Error(inputData);
log.Error(ex.Message);
return req.CreateResponse(HttpStatusCode.InternalServerError);
}
您可以在下面的附件中看到ResultCode为500,绿色状态...为什么? 我认为由于这个问题,Application Insight无法在错误/失败的请求查询中显示此数据.
You can see in below attachment, ResultCode is 500 with Green Status... Why? I think because of this issue Application Insight not showing this data in Error/Failed request query.
在Application Insight中找不到记录
No record found in Application Insight
exceptions
|where operation_Id == "c5b5a345-fa11-4356-b769-b34d1c6619e5"
| order by timestamp desc
| project operation_Id , timestamp
成功检查表示您的Azure Function调用是成功(=未引发异常)还是失败(=引发异常).
Success check denotes whether your Azure Function call succeeded (= no exception thrown) or failed (= exception thrown).
第一次调用时发生异常,因此函数调用未正常终止,因此出现红色选中标记.
At your first invocation, an exception occurred, so function invocation didn't terminate normally, thus red checkmark.
当您手动捕获并返回500时,就函数调用而言还是可以的-它完成并返回结果.
When you catch and return 500 manually, that's still OK in terms of function invocation - it finished and returned the result back.
函数运行时不遵循HTTP语义,规则对于所有触发器类型都是通用的.
Functions runtime doesn't follow HTTP semantics, the rules are universal for all trigger types.
Application Insight默认指标不会在失败的请求图中显示已处理的异常,开发人员需要针对已处理的异常进行构建查询,例如
Application Insight Default metrics won't show handled exception in failed request graph, Developer need build query for handled exception e.g.
requests
| where success == "False" and timestamp >= ago(7d)
| join kind= inner traces on operation_Id
| project operation_Id , timestamp, message, severityLevel
| order by timestamp, operation_Id
severityLevel :- 1 = Info and 3 = Error