allow data to be pushed to influxdb about cadvisor HOT 20 CLOSED

The plan as you mentioned is to have plugins for different storage
backends.
AFAIK we haven't explored influxdb in detail. Are you interested in working
on adding support for it?

On Fri, Jun 13, 2014 at 11:34 AM, Jonathan Chauncey <
[email protected]> wrote:

• How do you tell cAdvisor which influxdb nodes to use?
• Should it push the data in realtime or on a timer?

Im guessing eventually this will be the basis for a plugin setup for
allowing data to be pushed to other systems like ganglia and graphite.

—
Reply to this email directly or view it on GitHub
#39.

from cadvisor.

yeah I think im going to take a stab at it. Im definitely not a go expert but think I can get a working prototype on a branch for you guys to look at.

from cadvisor.

Definitely! Let us know how we can help.

from cadvisor.

Here's a basic shipper between the cadvisor and influx DB APIs: https://gist.github.com/erikh/fbfe6dadc938cf815435

I'll be working on this a bit as well. If you'd like to collaborate @jchauncey, please lmk.

from cadvisor.

+1 to plugins for different storage backends, we could put a @librato plugin together then :-)

from cadvisor.

@monnand started work to refactor the existing code to make it easier to add these drivers. So hopefully by next week you'll be able to add many different drivers for storage backends.

from cadvisor.

#58 is the last PR to add storage drivers into cAdvisor. Once we got this merged, I will work on other storage backends like influxdb and BigQuery (which is mentioned in #43)

from cadvisor.

On Jun 17, 2014, at 1:55 PM, monnand [email protected] wrote:

#58 is the last PR to add storage drivers into cAdvisor. Once we got this merged, I will work on other storage backends like influxdb and BigQuery (which is mentioned in #43)

Can I take influx? I can probably have it banged out later this week.

-Erik

from cadvisor.

@erikh that SGTM. Only question I have there is if it can be run within the same process as the rest of cAdvisor

from cadvisor.

On Jun 17, 2014, at 2:15 PM, Victor Marmol [email protected] wrote:

@erikh that SGTM. Only question I have there is if it can be run within the same process as the rest of cAdvisor

I’ll investigate this but it might not be possible. What would you suggest in lieu of that?

-Erik

from cadvisor.

I just wanted to keep us from depending on another process running, but if there is nothing we can do then we can go that route.

from cadvisor.

@erikh Thanks! You could take a look at the storage package and implement the StorageDriver interface. I probably will add a Close() error method to the interface, which seems useful.

from cadvisor.

Why isnt this configurable? Its memory by default and you can set a flag to push to influx.

from cadvisor.

@jchauncey Yes. I think that would be better. The only reason I didn't make it configurable is because we only have one storage driver now. Once we add more drivers, we should let users to choose which one they want to use.

from cadvisor.

👍 to configurable. For now we only have one option :D

With @monnand's StorageDriver PRs we should be able to take any of these easily

from cadvisor.

👍 to not restricting the implementation of this to a single service.

Being able to actively push to multiple distinct backend services, concurrently, is super desirable.

from cadvisor.

@erikh Do you have a patch we can collaborate on?

from cadvisor.

Unfortunately I have not had time to do this.

If you want to take it, go for it.

-Erik

On Jul 1, 2014, at 1:40 PM, Rohit Jnagal [email protected] wrote:

@erikh Do you have a patch we can collaborate on?

—
Reply to this email directly or view it on GitHub.

from cadvisor.

Working on this issue today. But failed to use the influxdb's Go library. Posted a message on their mailing list. https://groups.google.com/forum/#!topic/influxdb/FDU6ELiICGo

from cadvisor.

The initial version of this is done with @monnand's #74. Feel free to open other issues for details that may be missing :)

from cadvisor.