We are using licensed camunda 8 in prod but from last few weeks we observed that our operate loading data slow…we use operate to monitor prod processing. Even after processing is completed, and we verified the client application logs, still in operate it shows processing…it takes an hour or two extra to load all the data…
Is there a way we can improve this???
One suggestion was to add more operate pods with just exporter enabled but i was not able to find any resource related to it…
Hi @Bittu - are you experiencing this in SaaS or in a self-managed installation? The reason for it is because Operate (and the other components, like Optimize and Tasklist) use a cold, read-only data store for operations; the data needs to be exported from Zeebe then imported/indexed into that data store. There are options available to adjust how that process works, but it depends on whether you are using SaaS or self-managed.
If you are using self-managed, I would recommend checking out the available configuration options here. If you are using SaaS, because you have a license, your best - and fastest - option is to contact the Camunda support team.
(Edit - I should clarify that typically the export delay is in minutes, not hours, so there’s definitely room for improvement here!)
Hi Nathan,
we have self managed and we are connecting with camunda team and waiting for their response…but meanwhile we wanted to try few things to resolve this in dev since we are getting it in prod… camunda team suggested to add a env variable in operate related to post importer ignore missing data but i am not sure if this will help… we are also exploring to seperate excporter,importer and webapp from operate and have multiple importer but we are not sure if this will resolve our issue
@Bittu - this is likely a situation where there’s multiple possible causes, and therefore multiple possible solutions, and finding the right one will require some troubleshooting and testing. It depends on environmental factors (vCPUs, I/O throughput, RAM, etc.), as well as process factors (how many PIs at one time, how complex, how many tasks/jobs, how many variables, etc.).
we think so…but the issue is everything else is working fine…there is no issue in processing and neither pods resources are exhausted…same is for operate… only this one issue that when we have higher load then operate becomes slow and it takes around and hour or two extra to show everything in operate after all instances gets completed