0

We are using backstage for our internal developer portal and recently we are facing a very peculiar behavior. The nodejs application is randomly slowing down in a non-deterministic manner and degrading to a stage where the application completely dies down. It takes more than a min to load catalog/login via aad, a restart quickly fixes this issue only be creep back in after some time. We cant link any tangible reason on why such a thing can happen. We tried to do some basic profiling and found:

  • there are no apparent memory leaks
  • cpu usage is well under limits
  • List item

Has anyone come across anything similar? Any help to identify the root cause will be really helpful. We are struggling for the past couple of weeks to determine the root cause, we even tried to apply past releases but the issue seems to persists.

I can provide more details if needed.

TIA

Biswadeep Basu
  • 49
  • 1
  • 8

1 Answers1

0

How is your application deployed? Is the front end and back end all co-located with the database? Is your database the out-of-the-box SQLite in-memory one, or a separate Postgres DB (and if this, where is the DB?) Are you doing techdocs and if so, are you doing the out-of-the-box config, or the recommended one for production? Are your sources getting crawled when the slowdowns occur?

Having everything all co-located will make it challenging to sort out causes.

Frizbog
  • 11
  • 2