Since i'm using ELK version 5.3, may be is this the issue?
It's certainly possible there's a bug in 5.3. You're the only person I've seen report this particular error in quite some time though. It's not common, and is almost always related to the network stack, which is why I asked whether you're using a proxy, that's a really common source of this issue.
Can you please help in how to check out the request details in the "Automating Report Generation" section.
Oh, I'm really sorry, I meant to post a link to the Reporting 2.x docs. If you scroll down to the PUT _watcher/watch/error_report
request, you can see in the watcher definition the "http" section, with an object that look like this:
{
"content_type" : "application/pdf",
"request" : {
"method": "POST",
"headers": {
"kbn-xsrf": "reporting"
},
"read_timeout": "300s",
"url": "http://localhost:5601/api/reporting/generate/dashboard/Error-Monitoring?_g=(time:(from:now-1d%2Fd,mode:quick,to:now))&sync"
}
}
It's from this that you can get all the information you need to use curl. Specifically, it needs to be a POST request, and include the kbn-xsrf
header. You can ignore the read_time part, the flow changed from 2.x to 5.x. You'll get a response with another URL in it, which is the URL you need to hit to actually download the report and any information about the process. Then you basically just poll that URL until you get a 200 response, which means the report is finished being created, or has failed. This is the URL the Reporting management page in Kibana uses to server up the PDF or display the error message.
All that said, I don't expect anything different to happen here. When you click "Generate Report" in Kibana, it's basically doing the same thing you'd be doing manually with curl. From there, it becomes a task in a queue that uses Elasticsearch, but that's not the issue, it's trying to process the queue, but it fails to load Kibana so it can't even begin to create assets for the PDF.
Back to your setup, a single instance of both is the easiest use case, and it should work out of the box. I'm sad to hear it doesn't. Again, I have to ask again for more information about your setup, I haven't seen anything yet that would point to why you are seeing that error. Are these the same settings you used with the 5.2 instance? Have you changed anything else on the server that's running Kibana from when things worked on 5.2? Do you have any kind of firewall rules on the box hosting Kibana, in particular any that you've changed? Is this a container, or maybe a VPS, a real server that's running Kibana? If you're able to, if you run the 5.2 instance on that same box, does it still work?