Instead of relying on Koios and Ogmios, for those of use who run an Ergo Node index on, and Cardano db-sync, can we get support to self-host the API calls to be a watcher.

This would support independent and even more decentralized watching.

More specifically, which Koios api endpoints are required by the Rosen Bridge watcher?

I had 4 permits watching Cardano and three of them have gone missing. Could someone look into why when they get a chance please. Few others are reporting the same in Discord. Thank you and happy new year.

Currently the watcher configuration does not have ARM support, for devices such as raspberry pi's, orange pi's, or Mac Mini M1 (Asahi Linux Arch).

From the Telegram, apparently the postgres image does not have an arm build.

Could we provision and update for ARM support please.

As part of the setup process for a Watcher, operators are asked to set environment variables via a .env file and secrets such as their Ergo mnemonic and Koios AuthToken via local.yaml. This requires those secrets, particularly the ones in local.yaml, to touch the disk. This could expose users' secrets unintentionally, especially if they unknowingly try to push a git commit or pull request from their local operator repo.

In my understanding it is best practice to have these secrets read in via environment variable as opposed to having them hard-coded in a config file. I worked on a way to do this in my own setup, but it seems that the application logic would need to be configured to perform the substitution. For example...

network: cardano
ergo:
  type: explorer
  initialHeight: 1155237 
  mnemonic: ${ERG_MNEMONIC}
  node:
    url: https://node.ergopool.io
cardano:
  type: koios
  koios:
    authToken: ${KOIOS_AUTH_TOKEN}
  initial:
    height: 9669205
    hash: df10db0a29e19f2b66c5cef3eecb959ee28f30d676b4af44947fc02e25fc5c76
    slot: 110404676

...would be an ideal setup for this I think. But this would fail because the application would not know to perform the substitution.

Additionally, this would help deployment in a variety of other use cases such as infrastructure-as-code or container orchestration deployments, as it would be relatively straightforward to pass in the environment variables, git clone the repo, and then programmatically deploy without having to manually edit files.

Is there a way to implement something like this?