Distinguish between reporting probes and controlling probes #1043
Conversation
|
It would be good to know why |
2opremio
changed the title
|
We should also distinguish between reporting hosts and containing hosts, otherwise we get strange stuff in the UI like this container details pane:
(note the three hosts on the top). I will create a separate ticket for this. |
|
Its not go vet that catches that, its errcheck. |
| @@ -209,4 +209,6 @@ const ( | |||
| HostNodeID = "host_node_id" | |||
| // ProbeID is the random ID of the probe which generated the specific node. | |||
| ProbeID = "probe_id" | |||
| // ControlProbeID is the random ID of the probe which controls the specific node. | |||
| ControlProbeID = "control_probe_id" | |||
This comment was marked as abuse.
This comment was marked as abuse.
Sorry, something went wrong.
|
I'd remove the extra ControlProbeID, its redundant - ProbeID is only used from controls, and it was getting added where it shouldn't - by the host tagger, on container nodes the kubernetes reported is reporting. I would stop the host tagger from adding the ProbeID to all nodes that probe reports, and instead make the docker report add it to the nodes it controls. And update the comment on ProbeID constant. Does that make sense? |
|
I'm Tom and I approve of this change. |
Distinguish between reporting probes and controlling probes
Fixes #1040