Investigating failing medic-api

Hello all, We seem to be experiencing some unexpected down time with the CHT core application. This issue affects users logins or data syncs, as the server seems to not be taking in requests.

This issue seems to be resolved with a server restart that allowed a few uses login at a time. But within a period of 10 - 20 minutes, CHT starts to stop processing requests like before.

Have you experienced this, any possible lead would be of great help. Thanks in advance

screenshots
this login page for one of our vht’s remained for more than 5 minutes

after 12 hours

Cpu utilisation while user login is ongoing. This screenshot was taken after 8 minutes from attempted login.

Users who are logged can not sync to unknown

you can access the log files using this link

Environment

  • Instance: (eg: alpha.dev.medicmobile.org, etc)
  • Browser: Firefox, Chrome, incognito mode
  • Client platform: Windows, MacOS, Linux
  • App: api
  • Version: 3.6.0

Hi @atria

If this is indeed an instance running 3.6.0, I would suspect this: Replace underscore with lodash in replication · Issue #5942 · medic/cht-core · GitHub