Multiple instances monitoring on the same server

[myroslav]

From @slavikvin on May 7, 2015 11:17

Hi, I am running 3 couchdb instances on the same sever and trying to add munin monitoring features for all of them using your plugin. I tried to separate them by title, but my munin server have some collisions with that :( Can you recommend me something? Any helpful information would be appreciated.

Copied from original issue: kxepal#2

[myroslav]

From @kxepal on May 7, 2015 11:23

Hi,
Solution for you should be something like that:

1. Place couchdb_ munin script somewhere
2. Create symlinks against it in plugins.d/ directory like couchdb_instance_1, couchdb_instance_2 etc.
3. In conf.d/ create a file with configuration for the each:

[couchdb_instance_1]
env.uri    http://localhost:59841

[couchdb_instance_2]
env.uri    http://localhost:59842

...

and with additional options as well by your need.

Thought, didn't test such setup, but should works.

[myroslav]

From @slavikvin on May 7, 2015 12:25

Thanks for quick answer, just tried out your solution, the munin server cgi shows separate graphs for database fragmentation and database documents, but all the graphs are in single example , it seems that data from all instances writing to same rrds, except the database fragmentation and database documents which include database name in metric name. Any suggestion?

[myroslav]

From @kxepal on May 7, 2015 12:28

Hm, indeed it will because all the graphs will have the same name - I miss that moment. Then worth to fix that (:

[myroslav]

From @kxepal on May 7, 2015 12:29

Also, let's move this feature request to original repo: https://github.com/gws/munin-plugin-couchdb as it's the base one and I have commit bit there to handle your needs.

[myroslav]

@kxepal are there any news regarding this issue?

[kxepal]

@myroslav Thanks for the ping (:

Sorry, I had not a lot of time to handle this in back days. Will try to handle it in next few days. It'll also need me for CouchDB 2.0 support (not required, but pretty good addition for dev-cluster monitoring). Stay tuned!