metrics graph are empty v3.4.1

[chitenderkumar]

General information

• **OS Platform and Distribution **: centos7
• ElasticHQ Version: V3.4.1
• Elasticsearch Version: 2.4.0
• Python version: 3.6
• Browser Vendor and Version (if applicable): chrome

Issue Description

when we are trying to access the metrics graphs of a cluster. metrics graphs displayed empty.

kindly support us here, to make it work.

[royrusso]

Anything in logs?

[MHi0]

Hi Roy, I am facing the same issue on the following platform:

General information OS Platform and Distribution : CentOS 7 ElasticHQ Version: V3.4.1 Elasticsearch Version: 5.6.11 Python version: 3.6 Browser Vendor and Version (if applicable): Chrome and IE

Log is attached.

elastichq_metrics_graph_failure.txt

[sdesse]

Hello, Same issue with Debian.

elasticHQ is installed on a machine inside our network and published through a reverse proxy. I can see that the browser is trying to load a page that does not exist and gets a 404: https://oursite/socket.io/?EIO=3&transport=polling&t=MRoAXOo

The ElasticHQ publication URL is the following: https://oursite/somepath/elastichq/

OS Platform and Distribution : Debian 8 ElasticHQ Version: V3.4.1 Elasticsearch Version: 6.4.2 Python version: 3.4.2 Browser Vendor and Version: Chrome and IE

I don't see the index to store the data in elastic either. Thanks

[royrusso]

... so you don't see the .elastichq index created? Does the main dashboard have data on it?

[sdesse]

Hello, The .elastichq index is here and the application is working.

I have attached the screenshot:

Other strange behaviors include no auto-refresh and no persistence of the node list sort order and list size.

This is the content of the .elastichq index: { "_index" : ".elastichq", "_type" : "data", "_id" : "hqsettings", "_score" : 1.0, "_source" : { "index_name" : ".elastichq", "historic_days_to_store" : 7, "doc_type" : "data", "websocket_interval" : 5, "historic_poll_interval" : 300, "store_metrics" : true, "doc_id" : "hqsettings", "version" : 1 }

[royrusso]

So the metrics (websocket) approach is failing then. Did you deploy using docker or ?

[sdesse]

just duplicated the git repo in a directory, started the service and I have apache on the same host that publish the app

[royrusso]

Hmm... is Apache acting as a reverse proxy to ES?

[sdesse]

yes https://oursite/somepath/elastic/

[royrusso]

Ok. Let me see if I can reproduce. I vaguely recall similar issue before with reverse proxies and HQ Websockets.

[royrusso]

BTW, when you go to the metrics section, can you attach the log to this ticket? Just the part when you land on the metrics section. I need to see what the calls look like.

Make sure to cleanse anything you don't want public in the log.

[sdesse]

If I click the Metrics button there is no logs in the elasticHQ/application.log but if I load other pages there are. Only thing I get is this:

https://oursite/socket.io/?EIO=3&transport=polling&t=MRoAXOo

[dbapramod882]

Is this fixed.

[royrusso]

Should be fixed in develop. Next release should be soon after the ui issues are resolved. Upgrading the socket related libs seems to have fixed this.