Investigate issues for NRIS-EPD and NRIS-EMLI Importers

[Keegnan]

Problem Description The imports are taking up to 19 hours. They are coming up with status failed and are not sure if the records are writing or not.

• EPD is taking 9 hours and failing

• EMLI is taking 3 hours and failing

In order to (achieve some goal), (a system or persona) needs to (some action).

Solution Needs

• Find out why they are coming back with a status failed and why imports are taking so long.

Timebox

• About 2 days. If no answers by then we can reassess.

Outcome Details describing the outcome of the research

• Understand the cause and what steps need to be taken to resolve the import status and why taken so much time
• Understand if this is fixable by sustainment (within our domain of work)

Additional Context

• Review ticket NRPTI #1246 for more info.

[sggerard]

I believe both EMLI & EPD are taking a long time and have a failed status due to problems with the NRISWS API. It appears that the API becomes slow and starts throwing 500 errors late at night/early morning. I was able to replicate these errors in postman at 9pm.

It likely doesn't help if both our TEST and PROD environments CRON jobs end up overlapping.

Additionally the EMLI integration isn't actually processing any records at the moment due to the inspection sub types changing which will be fixed in ticket #1238

[sggerard]

Both Test and Prod show these errors only since replacing the COCOCOLA server with new hardware.

The main times these errors seem to occur is from 8-10pm and 3-7am.

[sggerard]

I have created a service desk ticket https://apps.nrs.gov.bc.ca/int/jira/servicedesk/customer/portal/1/SD-119888

[sggerard]

Kris Clarke has been investigating the issues raised in my ticket and has doubled the JVM memory of the NRIS-WS API. Will keep an eye on the logs and see if this has any performance improvement.

[sggerard]

After the memory increase we are no longer seeing any 500 errors!