I store global parameters for my flow such as a user credential, API server URL, alarm destination emails... into flows' environment variables. I have standard subflows that are re-used in many flows that can get their setting this way so I don't have to configure them one by one.
But I'm facing some problems with flow's environment variables (Node-RED 2.2.2 running as docker under AWS Fargate, data folder is persistant in an EBS storage)
- The most critical : sometime the password (which is an env of type credentials) resets itself to the default
__PWRD__value. When that happens, my flow becomes totally not-functionnal. I have even been blocked by the auth system because of too many attempt at night with PWRD as the password (I have no added a detector for that issue).
Other variables are untouched
I can't figure out the reason
I am 100% sure that the container has not restarted:
- I have a restart detector in my flows (a inject node set for single inject at restart that then push a message to a slack channel)
- no Fargate logs confirm a restart
Also I already have restarted willingly the container without loosing any env variable
- the annoying one : When a variable value is changed and deployed, the new value is not necesseraly taken into account or it may take some time (minutes, a lot of minutes).
Same, I haven't found anything replicable.
It happens, sometime, but I can't relate it to anything.
Have anyone faced any of those env variable issue ?
Thanks for any hint or advice on how to track that better.