Open jezzarax opened 9 years ago
The token is intended to be used to uniquely identify clusters.
At this time it is not used by pgantenna. There is much to do.
Are you interested in having a dashboard that displays info a for a fleet?
I have several postgresql clusters I want to monitor on a single pgantenna. I'll try to dig into the code :)
@jezzarax a connection to pgantenna begins with the "handshake" command which is generated in handshake.c in pgsampler.
When the command is processed in pgantenna, an instance variable is set @token in lib/antenna/commands.rb.
Thus every connection received and created will have a @token identifying which cluster it is. Unfortunately supporting multiple clusters requires saving the cluster_id in a table, so that it can be associated with all the metrics collected. This results in a much more complicated implementation on the server side. For example, users in the single cluster use case would have to scope queries by cluster_id. This complexity is why it was left out for an initial draft to gauge interest.
Please ask if you have any questions.
Hi! Maybe I missed it in the documentation or database structure, but is there token column that will allow me to separate different clusters in graphs?
More general question, is it possible to collect data from several clusters into one pgantenna instance and separate collected data by cluster?
Thanks!