Report startup failures in the app Log Stream

An example surfacing logs properly is this issue where the worker crashes and the host restarts, but this information is not logged to users as errors. The host process was not able to start the worker process and it logged the startup exception as information

Customer comment: “the startup failures are being reported to Insights (and not appearing at all in the Log Stream of the app). It should’ve been logged as an error and it should also have appeared in the Log Stream.”

// Scope: clusters/wawscus/databases/wawsprod
let st = datetime(2021-12-09T13:26:43.8280758) - 5m;
let et = datetime(2021-12-09T13:26:43.8280758Z) + 5m;
FunctionsLogs
// | where PreciseTimeStamp > ago(30d)
| where PreciseTimeStamp between (st..et)
| where AppName =~ "testesgerais-dev"
| project-reorder PreciseTimeStamp, Summary, FunctionName, Details, HostVersion

User story: Log startup failures in the app Log Stream 

[fabiocav]

Since this is in Application Insights today, we should ensure that:

• Initialization errors have the appropriate level in app insights
• We log to diagnostic events, so when that comes online, that information is available in the portal

Let’s repro with an initialization erros (perhaps using the setup referenced by @liliankasem) and validate the above requirements.

[fabiocav]

We need to continue investigation here.

@satvu has the repros, so our next step should be to identify if there’s an issue with the host (meaning, the worker wrote the failure details to stdout/error, and the host did not appropriately log them), or if there’s an issue with the worker implementation.