In doing some performance testing, @nessiLahav found that server-side errors can still return the default Rails error page. We should ensure that Rails does not do this.
Steps to Reproduce
Trigger a 500-level server error, either through inserting a fake exception or causing a Sequel:PoolTimeout by applying load as @nessiLahav did.
For instructions to reproduce, follow one of the failure scenarios here
Note: Any Postgres errors for examples (too many client, pool timeout) will reproduced the issue
Look at the error page.
Expected Results
Production mode error page should indicate an error occured but should not give error details. Development mode error page should give the full stack trace of the issue.
Actual Results (including error logs, if applicable)
The default Rails status page appears.
The rest I'll leave to @nessiLahav to add more details about his setup:
We should start the investigation by simulate Postgres errors by UT or in debug mode, in different flows in the code (not just HFT) to understand the behavior.
Summary
In doing some performance testing, @nessiLahav found that server-side errors can still return the default Rails error page. We should ensure that Rails does not do this.
Steps to Reproduce
Expected Results
Production mode error page should indicate an error occured but should not give error details. Development mode error page should give the full stack trace of the issue.
Actual Results (including error logs, if applicable)
The default Rails status page appears.
The rest I'll leave to @nessiLahav to add more details about his setup:
Reproducible
Version/Tag number
Conjur version: 5.8.1
Environment setup
DAP deployment setup: 1 DAP Master + 1 synchronous standby + 1 async standby Conjur Image: CentOS 7 Cloud: AWS
Additional Information
We should start the investigation by simulate Postgres errors by UT or in debug mode, in different flows in the code (not just HFT) to understand the behavior.