[
https://issues.apache.org/jira/browse/NIFI-4344?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Denes Arvay resolved NIFI-4344.
-------------------------------
Fix Version/s: 1.14.0
Resolution: Fixed
housekeeping: marking this as resolved with 1.14.0 fix version as this has been
already committed for a while.
(https://github.com/apache/nifi/commit/0c748a5a2b366990c544e0580ca8626f0b9e3ed1)
> Improve bulletin messaging with exception details
> -------------------------------------------------
>
> Key: NIFI-4344
> URL: https://issues.apache.org/jira/browse/NIFI-4344
> Project: Apache NiFi
> Issue Type: Improvement
> Components: Core Framework, Core UI
> Reporter: Pierre Villard
> Priority: Major
> Fix For: 1.14.0
>
> Time Spent: 2h 20m
> Remaining Estimate: 0h
>
> In some environments it is not possible/allowed to access the NiFi nodes (and
> consequently the log files). In such a situation, when developing a workflow,
> the only option to understand what could be wrong with a processor is to look
> at the bulletins.
> However, the bulletins only contain the message of the main exception. Quite
> often, the very useful messages to understand what is going on are located
> into the wrapped exception. At the moment, the only option is to go into the
> log files to get the messages. Even when access to the log files is
> authorized this can unnecessarily slow down workflow development.
> Ideally the full exception should be stored with the bulletin. However
> granting access to a full stack trace can raise security concerns, and
> besides, the bulletin tooltips are not designed to display large amount of
> data.
> Additional data would be provided in the bulletins view only. Few options are
> available:
> - display the full stack trace in the bulletins view but only for users with
> restricted component authorization.
> - display an enriched message with the concatenation of the detailed messages
> of the stack trace in the bulletins view (with or without restriction).
> Adding an example to illustrate the problem: with ConvertRecord, in case
> something is wrong with the schemas for the input/output data, the error is
> handled with:
> {code:title=ConvertRecord.java|borderStyle=solid}
> } catch (final SchemaNotFoundException | MalformedRecordException e) {
> throw new ProcessException("Could not parse incoming data", e);
> }
> {code}
> In this case the bulletin will contain very little information.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
