I use Central Vault setup for multiple kubernetes clusters and usually it works fine.
But occasionally pods for cron jobs are failing and the logs are
failed to create vault clienttimeout [10s] during waiting for Vault token
I use GCS as the backend for Vault.
This behaviour is not consistent and so I am not able to understand what it can be.
Please if someone can help.
H’m. Can you share a bit more about what these pods are doing? I wonder whether they’re trying to manipulate the VAULT_CLIENT_TIMEOUT environment variable.
The pods are doing some general processing after reading a few messages from the DB. They do not do any manipulation on the Vault side.
Also out of curiosity, will it be better if we use postgres as the backend instead on GCS.
Is there a latency issue when using GCS as backend.