bridge: remove all supervisor.SignalHealthy calls

[leoluk]

Stack from ghstack:

• solana: add uncommitted Cargo.lock files #68 solana: add uncommitted Cargo.lock files
• bridge: migrate cmd/ to cobra #67 bridge: migrate cmd/ to cobra
• Remove outdated TODO comments #66 Remove outdated TODO comments
• bridge: refactor p2p logic into pkg/p2p #65 bridge: refactor p2p logic into pkg/p2p
• bridge: remove all supervisor.SignalHealthy calls #64 bridge: remove all supervisor.SignalHealthy calls
• bridge: refactor processor logic into pkg/processor #63 bridge: refactor processor logic into pkg/processor
• node.proto stub and dependencies #62 node.proto stub and dependencies
• devnet: use real account and nonce for send-lockups.js #61 devnet: use real account and nonce for send-lockups.js
• bridge: bypass p2p for our own signatures #60 bridge: bypass p2p for our own signatures
• bridge: verify LockupObservation signature #59 bridge: verify LockupObservation signature
• bridge: correctly calculate 2/3+ majority #58 bridge: correctly calculate 2/3+ majority
• IDE helper scripts for logs and lockups #57 IDE helper scripts for logs and lockups

Supervisor does not back off tasks that failed in a healthy state.

There are a couple places where we rely on supervisor for
application-level backoff, so we always want back-off. The distinction
is meant to enable runnables to implement their own specific back-off
logic, which we don't, so we can safely ignore it.

Fixes #37