This message was deleted.
# opal
s
This message was deleted.
r
I’m almost sure that no, only the dict itself.
Also if you’re using k8s, you should escape the value.
a
Which value?
r
I mean that you should escape the value you are passing in the
OPAL_DATA_CONFIG_SOURCES
env.
a
I think I’m following. I’ll probably have the Helm chart format this value somehow.
r
yeah sure, let me know if you need help with that, escaping with helm can be frustrating sometimes 😉
a
I’m more worried about the functionality of my data fetcher than anything else.
At least for now. 😬
o
You got this 😉 💪
a
I’m trying to write my data fetcher to be as generic as possible. As I understand the config will come into the OPA system and basically define what to fetch. But what I’m figuring out now is how that comes into play with
_process_
. Because I’m not sure if I can find some clever way to allow an end user to supply their own ‘transform’ logic. Maybe by including a
func
as an input to config?
o
Mmm - that does sound super generic. Who’s the end user in this scenario? A dev on your team?
a
Well, if I can write the perfectly generic data fetcher… the public can be the end user. But to complete what I need for work it’ll be other devs in my department.
o
Depending on how complex you need the transform to be. You can as go as far as including a script to execute as part of the fetcher config in the update event. But you need that to be super secure - to avoid code injections You can try something like EBPF, or if you need just data extraction - maybe something like regex or glob pattern would be enough
Or maybe this a Javascript sandbox based on node - with a Python wrapper. https://pypi.org/project/node-vm2/
a
I might be over-complicating this for our case.
o
Could be too 😉