[GitHub] GabrielBrascher edited a comment on issue #3078: Add influxdb to statscollector

[GitBox <gi...@apache.org>]

GabrielBrascher edited a comment on issue #3078: Add influxdb to statscollector
URL: https://github.com/apache/cloudstack/pull/3078#issuecomment-444623455
 
 
   Hi @bwsw,
   Thanks, I appreciate your suggestions. To avoid misunderstandings, I would like to start by saying that this implementation focus on adding support to InfluxDB (**as a driver**) in the StatsCollector that we already have in ACS. Your proposal to add a driver in StatsColletor to send data to the EventBus is for sure interesting, but it is out of scope here.
   
   Now, addressing your points:
   This PR enables ACS to send StatsCollector's data to InfluxDB (we are just adding a driver to support the integration of InfluxDB and ACS). Additionally, as you can see, StatsCollector works fine but needs a huge refactoring, extracting code, adding tests, documenting, and so on; this implementation is a starting point for all of that. Even in such a "simple" InfluxDB implementation, it took hundreds of lines to fit in the StatsCollector codebase right now.
   
   Regarding the 3rd party issue; as you mentioned, we are a small community and cannot afford to waste time, resources and have pieces of code that nobody will know how to test, use, and extend. That is indeed why I do not see any problem on using the InfluxDB 3rd-party component; with a few well-documented lines, it allowed me to create an InfluxDB connection and write the stats (again, I am only introducing a driver here to enable ACS to communicate/send data to InfluxDB).
   Regarding the high-availability, not all use cases require such feature. I do see the value of your suggestion and agree that it is an interesting feature to add to CloudStack; However, I do have a problem with the idea of forcing all users to prepare a redundant environment to support such HA (multiple nodes running InfluxDB + the Kafka environment). Both approaches can coexist and I encourage you to create a PR with such feature, I will gladly help to review it.
   
   Moreover, are you aware of https://github.com/influxdata/influxdb-relay? It might be interesting to check it. Another option (later!) is to implement a Telegraf driver; it is a stats proxy which can send stats to InfluxDB, Graphite, ElasticSearch, etc. It would work with Kafka as well.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services