Add the option to automatically install TICK scripts from a specified directory#8
Add the option to automatically install TICK scripts from a specified directory#8devth wants to merge 2 commits intoinfluxdata:masterfrom
Conversation
kapacitor/values.yaml
Outdated
| ## tick_dir directory, if enabled. | ||
| ## | ||
| install_tick_scripts: | ||
| kapacitorURL: http://kapacitor:9092 |
There was a problem hiding this comment.
Isnt this just {{ template "fullname" . }}.{{ .Release.Namespace }}:9092?
| data: | ||
| init.sh: |- | ||
| #!/usr/bin/env bash | ||
| export KAPACITOR_URL={{ .Values.install_tick_scripts.kapacitorURL }} |
There was a problem hiding this comment.
Maybe export KAPACITOR_URL={{ template "fullname" . }}.{{ .Release.Namespace }}:9092
|
So you are planning to store your alerts in configmaps? I'm not sure how I feel about that. Could you potentially have a different method for storage? Github/lab? |
| name="${name/.tick/}" | ||
| kapacitor define $name \ | ||
| -type batch \ | ||
| -dbrp telegraf.autogen \ |
There was a problem hiding this comment.
Do we have a better solution to hardcoding the type and dbrp of the task? In the past I have parsed the type and dbrp values from comments at the top of the TICKscript. Part of the work in influxdata/kapacitor#1313 will be to standardize that behavior.
For example
// kapacitor:type=stream
// kapacitor:dbrp=telegraf.autogen
or
// kapacitor:type=batch
// kapacitor:dbrp=mydatabase.autogen
Something along those lines is what I was thinking, but that will probably not be the final result of influxdata/kapacitor#1313.
Technically it is possible to have multiple dbrps values but you can probably leave that out for now as it is uncommon.
There was a problem hiding this comment.
For my purposes I don't (yet) need it to vary, which is why I hardcoded it. I figured I could bake it into the dir structure if I needed to parameterize it someday, e.g. $TICK_DIR/telegraf/autogen/foo.tick. If you want to bake it into the TICK script itself and parse it out in kapacitor, sounds good to me 😁.
|
@jackzampolin in my usage, the alerts are stored in git as Helm templates. ConfigMap is just an easy way to get them mounted into a container without having to modify a Docker image to bake them in. The current method of specifying |
3 participants
As requested in influxdata/kapacitor#1313