avadodserver / aip-avado-improvement-proposal Goto Github PK

Right now it's impossible to tell which of the running containers is stressing the RAM/CPU the most. A more detailed output - perhaps even a grafana dashboard native to the avado itself - would help with debugging chaotic binaries.

It should be possible from the UI to:

• flush the dns cache
• remove and reload the IP lease
• do other IP-related operations

This is especially important for people with failover internet connections. A network cache can interfere with getting a new IP address from another router, and the node can experience prolonged downtime. A forced ip renewal would speed this recovery up.

Packages currently come with ports predefined. It would be useful to be able to open them and re-run a container with those ports, as if the container was run with -p 12000:12000. Actual example use case is Prometheus on both Kusama and Prsym beacon node - ports 9615 and 8080 respectively.

The link in the Video tutorials in the footer of the https://ava.do page seems broken.
Please replace with this link:
https://www.youtube.com/channel/UC6OLIepQMLmN-49FW4d1B-A

Thank you.

Certain packages (primary) will only work if other packages (secondary) are already installed and running.
Example: ENS resolver is a crucial secondary package for any primary package using the wizard, as the wizard is an .eth link

To ensure that users know about these dependencies, they should be informed when installing the primary package. If possible the installation process should check for the secondary packages and automatically install them, if they are not running.

When opening ports of the various containers, make sure both TCP and UDP are open. This is necessary for Ethereum to have Discovery v5 working.

The dashboard should be accessible through a public IP of the internet the node is connected to, for everyone who wants to manage their node remotely and is finding the Avado bridge unreliable or unreachable. The public availability of the dashboard should be HTTP-basic-auth protected and a toggle in the settings, so not open by default.

Sometimes direct access to data folders of the running containers is necessary. For example, extraction of keys, or a renewal of network keys to recover a node's network reputation. One should be able to see the directory structure of a container's data folder in the UI, and be able to explore it, delete, download, upload, and create files.

For safety, have a timeout of 10 minutes after changing. If the user does not log in using the new password and confirm the selection, the password should revert back to avado123.

Right now, when a critical update is issued, one has to wait for the team to update the container and provide a new binary. In validation environments this is especially dangerous. Users should be able to auto-dockerize and run the latest binary available on the release page of a project. Whether this should be accomplished via a direct run (no docker) or some kind of docker wrapper, I'm not qualified to speculate.