TriggerTopicAsync slows down
Hi novu-Community,
first of all congrats to the relaunch of the documentation - looks awesome!
We have self-hosted novu in our Kubernetes-Cluster on Azure. During setup we recognize that the helm chart, that you have in your Git is not published, so we copied it over (would be helpful for new customers, that the helm chart is published somewhere)
After deploying it to our cluster we did some testing and measured time of the
_novuClient.Event.TriggerTopicAsync
call and saw that the time increased from a few milliseconds to seconds and even more then 3 seconds sometimes. As this is not acceptable for our use case, I ask now myself if we have did something wrong or where the isuse could come from.
We deployed the MongoDB you configured in your helm chart to our cluster and we are right now on novu version 0.1.6
which is a bit older, but before upgrading we want to know where the performance issue comes from.
Would be happy to get some tips from your side, where the reason for this could be. We can also schedule a call, where we have a look together.
Best,
Sebastian4 Replies
Hi @sebastiankuesters! From what version to what version are you upgrading and saw the performance degregation?
Hi @dimagrossman we have never upgraded, we deployed the version 0.1.6 once and started to test there. During testing we detected this issue.
Just to clarify, it's 0.16.0 right?
Are you also running a dedicatd redis cluster for caching (not for queueing)?
And another question, have you verified that you have the proper indexes created inside your mongodb cluster?
Sorry, my fault - 0.1.6 was helm chart version... My novu version is 0.15.0
We are using Azure Cache for Redis as Redis cache. MongoDB indexes I have not checked - need to do that.