Gradual increase in memory usage

  • Is this your first time deploying Airbyte?: No
  • OS Version / Instance: Ubuntu - 20.04
  • Memory / Disk: 64GB/ 200GB
  • Deployment: Docker
  • Airbyte Version: 39.28

When we start Airbyte, the memory usage of airbyte-worker is around 1.5GB. Memory usage increase as time passes. Currently, it uses around 4GB. Is this expected? How to handle it efficiently?
PS: I know that we can limit the resources of each of the containers but wanted to understand what’s causing the memory increase to go up in the first place.

Did you create more connectors & connections or keep the same as the begining?

@marcosmarxm we have the same number of connectors and connections. The memory usage went by 300mb since the time I asked the question. So it’s now almost 4.3GB. Memory usage won’t come down even if there is a long after jobs stopped running.

And airbyte-server memory usage is also slowly increasing but at a slower pace, 5-10MB per day.

Do you mind sharing the output of docker stats ?

Here

Sorry the delay here Prudhvi, did the memory increase again or become steady now?

It is increasing, it crossed 9GB as of now. Current screenshot

Seems potentially related to Resource leak in airbyte-worker

Pudhvi and Adam I request the engineer team to check for what is the expected resources needed and if maybe a bug exists.

@marcosmarxm just wanted to check if there are any updates on this.

And here is current resource usage. It keeps on increasing:

The team is investigating a possible issue with memory leaky in the logging system. Until a solution is merged the workaround is to restart airbyte-server and airbyte-worker containers. I created the issue to track the issue: https://github.com/airbytehq/airbyte/issues/16652

1 Like