Better error handling

[tristanlatr]

NitroSession should handle correctly SIEM errors :

Currently, calls to NitroSession.esm_request that raises requests.HTTPError errors can raise NitroError if the error is unhandled.
See docs : https://mfesiem.github.io/docs/msiempy/index.html#msiempy.NitroSession.esm_request

• Unhandled 400 errors :

  • ERROR_InvalidFilter (228)
  • No such command
  • ERROR_Internal (-1)
  • ERROR_JEC_ResponseNotAvailable (3503)
  • Cannot deserialize value of type com.mcafee.siem.api.data.query.EsmTimeRange
  • Cannot deserialize instance of java.lang.String out of START_OBJECT token
  • Both customStart and customEnd properties must be set when running a query with a custom time range.
  • Cannot construct instance of com.mcafee.siem.api.data.alert.EsmAlertId
  • Parameter "xxx" cannot be missing
  • All other errors

• Unhandled 500 errors :

  • All

• Connexion errors -> Raises !

  • Read timed out
  • Max retries exceeded

[tristanlatr]

I've tried to implement an invalid session error handler. It re-logins the user, then it sleeps 0.2s and retry the request. If it still run into an error, It returns a NitroError

[tristanlatr]

The method NitroSession.esm_request currently handles the following errors :

• 'ERROR_InvalidSession', 'Not Authorized User', 'Invalid Session' or 'Username and password cannot be null' -> Authentication error, handled by re-login and retry request.
• 'ERROR_IndexNotTurnedOn', 'ERROR_NoData','ERROR_UnknownList', 'ERROR_JobEngine_GetQueryStatus_StatusNotFound' -> Data unavailable error, handled by simply returning a empty list
• All other errors will be retried one time and if falling again will raise NitroError

[tristanlatr]

The current behaviour for NitroSession.esm_request() is to automatically retry once the request if it failed. If it's a login issue it will re-login first.

The current behaviour for EventManager.qry_load_data() is the same: retry once the request