Dear all, sometimes we are having some slight delays in getting everything on UI (like up to 5 seconds), or when navingating through Datasets. It does not always happen, sometimes it is instant. Is there a way to raise the performance by allocating more resources to the containers, etc.?

Depending on the volume of entities and aspects you have, it will be necessary to vertically scale your elasticsearch cluster. But It's just an assumption. Did you try to enable the search cache?

How do I do that?

How did you deploy your instance? Helm chart?

I've deployed using quickstart and localized docker containers.

If you’re having trouble with quickstart, please make sure you’re using the most up-to-date version of DataHub by following the steps in the quickstart deployment guide: https://datahubproject.io/docs/quickstart/#deploying-datahub. Specifically, ensure you’re up to date with the DataHub CLI:

python3 -m pip install --upgrade pip wheel setuptools
python3 -m pip install --upgrade acryl-datahub
datahub version

Hmm... Understand. Did you fork the project? You can enable the cache just by adding the 'SEARCH_SERVICE_ENABLE_CACHE=true' environment variable on GMS docker container. You can rewrite this dockercompose file: https://github.com/datahub-project/datahub/blob/master/docker/quickstart/docker-compose.quickstart.yml Change the elasticsearch memory limit: https://github.com/datahub-project/datahub/blob/master/docker/quickstart/docker-compose.quickstart.yml#L117 And even set how many cpus to use. Like "cpus: 0.5"

If you’re having trouble with quickstart, please make sure you’re using the most up-to-date version of DataHub by following the steps in the quickstart deployment guide: https://datahubproject.io/docs/quickstart/#deploying-datahub. Specifically, ensure you’re up to date with the DataHub CLI:

python3 -m pip install --upgrade pip wheel setuptools
python3 -m pip install --upgrade acryl-datahub
datahub version

I did not come back to this, but doubling the elastic search memory to 4g actually helped with the performance issue.