Update kibana works without any effort deploying the python catalogue template a starter cluster and on "pro-us-east-1". Its broken on "pro-eu-west-1" for old and new projects and using the templates seen to work elsewhere.
Update: Here is a screen recording of the problem https://www.useloom.com/share/70469210881a46b69a1914462f8cc7ad
I have setup a process on OpenShift Online Pro using the official openshift/php:7.0. The health check is running so it is logging:
10.131.30.1 - - [08/Apr/2018:16:33:43 +0000] "GET / HTTP/1.1" 200 1023 "-" "Go-http-client/1.1" 10.131.30.1 - - [08/Apr/2018:16:33:53 +0000] "GET /health.php HTTP/1.1" 200 2 "-" "Go-http-client/1.1"
If I click on "View Archive" it takes me to a Kibana 4.6.4 dashboard. I see a "Warning No default index pattern." The Kibana documentation and prior experience tells me that Kibana should let me browse the fields to configure the indexing. Poking around I haven't been able to get any data to show on Kibana and it seems that there isn't any data being found in elastic search or that Kibana isn't connected to the correct log index. How do I bootstrap this? Thanks!
To reproduce on openshift online pro:
ab -c 5 -n 50000 <route> to try to force a flush to kibanaI tried the same steps on OpenShift Online Starter and Kibana gives the same Warning No default index pattern. You must select or create one to continue.
Update:
On one cluster logs.pro-us-east-1.openshift.com Kibana doesn't load with a JavaScript error. It should be showing a project called tiller but Kibana crashes out with an error referring to a different openshift project legalknowledge. Here is a screenshot where the top left it shows the demo project name created using the above steps laravel-mysql-persistent yet the browser javascript error is talking about a elastic index for a deleted project id: project.legalknowledge.e60761e4-3b39-11e8-b27b-12b5519f9b58.* https://www.dropbox.com/s/0pcgu5yhsw2yi8m/wrong-project-openshift.png?dl=0
Update: same problem on eu-west cluster where it gives the js error but has the correct project name.
The upshot was that the cluster admins had to delete some old elastic search indexes and provision more space.
So the technical answer is that you shouldn’t need to do anything to get “view archive” working and should raise a support ticket.
