Add add Automated Dashboard for Kubernetes Node metrics #64
Conversation
AppSignal for Kubernetes extracts node metrics for each node running in a Kubenetes cluster. This automated dashboard displays some of those metrics in an automated dashboard named "Kubernetes Nodes".
jeffkreeftmeijer
requested review from
matsimitsu,
roytomeij,
luismiramirez and
unflxw
jeffkreeftmeijer
changed the title
| "name": "node_fs_inodes", | ||
| "fields": [ | ||
| { | ||
| "field": "GAUGE" | ||
| } | ||
| ], | ||
| "tags": [ | ||
| { | ||
| "key": "node", | ||
| "value": "*" | ||
| } | ||
| ] | ||
| }, | ||
| { | ||
| "name": "node_fs_inodes_free", | ||
| "fields": [ | ||
| { | ||
| "field": "GAUGE" | ||
| } | ||
| ], | ||
| "tags": [ | ||
| { | ||
| "key": "node", | ||
| "value": "*" | ||
| } | ||
| ] | ||
| }, | ||
| { | ||
| "name": "node_fs_inodes_used", |
There was a problem hiding this comment.
Let's use tags for different states like free and used instead of reporting them as different metrics. We do this for other (host) metrics too. It would help that we don't have to show the full metric name then for every line in the graph, freeing up valuable space in the hover box.
For example:
- Metric name:
node_fs_inodes - Tags:
state- Values:
freeused
- Values:
I also see this in some other graphs in this dashboard. We should update those as well.
There was a problem hiding this comment.
I'm trying to keep this dashboard as close to what's reported from Kubernetes as I can. I think this is a good idea for the future, when we know what we'd like to report exactly, but let's get this out of the door and get users to try it first.
| { | ||
| "title": "Node CPU Usage", | ||
| "description": "node_cpu_usage_nano_cores", | ||
| "line_label": "%name% %node%", |
There was a problem hiding this comment.
We shouldn't include the metric name unless there are multiple metrics in a graph, and for those graphs we're better off using tags on one metric. See my other comment.
| "line_label": "%name% %node%", | |
| "line_label": "%node%", |
| "visuals": [ | ||
| { | ||
| "title": "Node CPU Usage", | ||
| "description": "node_cpu_usage_nano_cores", |
There was a problem hiding this comment.
These descriptions don't explain anything to me. Let's remove them if it's just for testing.
There was a problem hiding this comment.
Or if we can do human-readable descriptions, let's do those instead!
There was a problem hiding this comment.
These are the internal metric names, so they'll probably make sense to Kubernetes users.
There was a problem hiding this comment.
I can't the names of these metrics anywhere, so I wouldn't be so sure that it's clear. The same way I can't find that you're using the kubernetes metric names, like you say here. In the library I see it being mapped from a JSON struct that doesn't use the same naming: https://github.com/appsignal/appsignal-kubernetes/blob/0b3f39d65ba99622ab3e647e8e4c012ee944baca/src/main.rs#L97-L167
Do you have any links to docs or source code that mentions these metric names?
|
I'm not sure what I should review :) |
There was a problem hiding this comment.
If this is just a quick test with Jim and select users, then this is fine.
[But then, maybe we don't need a magic dashboard at all? Maybe we can just give them a JSON dashboard definition that they can manually import into AppSignal?]
Otherwise, what @tombruijn said.
And also maybe we could consider prefixing these metric names with something like k8s_? Otherwise it might be confusing with the Node.js Heap Statistic metrics (there's no overlap, though -- those are nodejs_, not node_)
|
Accidentally merged, reverted and continuing in #65. |
We're adding a dashboard to give users a bit of an easier time getting their Kubernetes setup working. This is mostly an example dashboard at this point, but it saves users from having to write their own.
You're right, this could be a bit confusing. I'm putting this on the list for later, as we're going to add more metrics that also warrant namespacing (think pod metrics, for example). I'll pick this one up then. |
AppSignal for Kubernetes extracts node metrics for each node running in a Kubenetes cluster. This automated dashboard displays some of those metrics in an automated dashboard named "Kubernetes Nodes".