Recently one of our client has asked us to provide a solution for an issue.
The issue is, the production support team is getting hundreds of emails when there is an exception in message flows by. They keep receiving the emails until the issue is resolved.
By receiving hundreds of emails for the same issue, the support team
- Inbox gets full
- Missed some important emails
- Unable to concentrate on other issues
- Need to clean up inbox frequently to get other emails
- Gets frustrated by seeing so many emails
The existing exception handler sends email alerts for every exception and this needs change now. The change required is
- Do not send hundreds of email alerts
- Send only for the first occurrence. For the subsequent occurrences, store the error message in memory and send a summary of errors after N minutes which has to be a configurable one
- The summary email should contain
- Message flow name
- Time of first occurrence
- Time of last occurrence
- Error message
- Occurrences count
- Above information must be in HTML table format
Initially we thought it’s very simple to implement but later realized that
- Storing table kind of structure in memory is complex
- Each flows error message must be maintained unique and needs to maintain the subsequent error details in memory
- If error is resolved before N minutes configured for summary email, still had to send email with summary details.
After brainstorming with the team the solution has been implemented using Java compute node API and global cache which worked perfectly and client was very happy.
Are you looking for same or similar solution? We exclusively focus on IBM Integration Bus Development, Support and Training. Feel free to contact us at email@example.com.