Add telegraf-ds (imported and updated from old tick-stack charts) #16
Conversation
nsteinmetz
changed the title
|
@nsteinmetz @rawkode Would it be too involved to configure the current telegraf chart such that it could be deployed as either a deployment or daemonset? |
|
It should be doable, the issue I see is more about readability and expectability. And maybe there could be some deployment issue if you try to deploy the DS and and the Deployment in the same namespace. Having a single chart for telegraf would mean that you must explicitely set what kind of Deployment or DS you expect and related plugin are not the same too. So the targeted chart may be hard to read and use. DaemonSet use case is more on how to monitor nodes and what they contain whereas deployment is more to monitor pods by themselves. So as use cases are not the same, I think it's better to have duplicated code and two different charts to ease readibility and understanding instead of a single one. |
|
100% agree with @nsteinmetz Lets not merge these, although they're both Telegraf's; they very different use-cases. I'm testing this chart now 👍 |
This chart does not allow tweaking the input configuration. It is for node and Kubernetes monitoring. Use the `telegraf` chart if you need to configure the inputs
… they want to store metrics
feat(telegraf-ds): opinionated telegraf-ds deamonset
…e fact that inputs cannot be customised.
Add telegraf as daemonset - closes #1