- Functionality: Acts as both a command-line interface and a service daemon, compatible across Linux and OSX platforms. Windows may be supported in the future.
- Management: Facilitates minimal validator and pool configuration, leveraging a 'manager' account for transaction signatures. The included UI is the preferred method of management.
- Participation Key Management: Automates the creation and renewal of participation keys to maintain pool activity and online status.
- The Reti Node Daemon Is a combination CLI / Service daemon that will run on Linux / OSX / Windows and which node runners should run as a background service.
- Each node daemon must have access to a 'manager' account hot-wallet which it can sign transactions with. This manager account can be switched out by the owner of that validator to a new account at will if there is a compromise.
- The only accounts that can ever remove user funds are stakers removing only their (compounded) balance. The 'epoch update' call is made by the daemon for all pools, all aligned by the epoch time. This epoch update is what initiates the payout to the validator for their commission as well as updating the balances of all stakers. Any account can trigger this contract call, as it allows anyone to update eligible rewards even if the validator goes offline somehow.
- On each node, it will monitor the staking pools defined and automatically create short-lived (1 week) participation keys with that nodes algod instance.
{% hint style="warning" %} Participation keys for pools assigned to a node are only created once stake is present in the pool ! {% endhint %}
- The participation keys will be monitored for expiration and new keys will be created in advance so that its always online.
- As participation keys are created, the paired staking pool will be instructed via the 'manager' to issue a transaction to go online against that participation key.
- The node daemon will likely provide a number of prometheus compatible metrics for scraping into compatible agents (staking amounts, etc.) but broader monitoring will be best handled independently.