Scalability, Performance and Robustness #187 Milestone 2: Reliability Milestone (RM)

Milestone 2: _(Reliability Milestone (RM))_

• Focus: In order to reach consensus on a single transaction, Canton requires sufficient responses of confirming nodes. While a response may be submitted through any sequencer, and it may be submitted multiple times to reduce failures, such resubmissions introduce additional latencies. The given milestone will improve the liveness signal of a sequencer and the sequencer selection algorithm such that only healthy sequencers are used, while at the same time create the necessary data in order to monitor the quality of service by individual sequencer operators.
• Estimated Delivery: Sept 1st, 2026
• Deliverables / Value Metrics:

- Write signal health check and connection filtering, such that submissions are sent to a sequencer which will not respond with SEQUENCER_OVERLOADED due to internal processing delays. - A new set of health metrics exposed by the validator, segregated by sequencer, detailing the availability of individual sequencer nodes for write and read, including associated failure rates. - A public website where the current observed health metrics of a list of Super Validators can be published to. - A publishing feed of our operated nodes to this website.

| Milestone | Code | Payment | Acceptance / Trigger | |-----------|------|---------|----------------------| | 2 | RM | 1,080,000 CC | Upon committee acceptance of deliverables |

_Originally posted by @pedrodneves in https://github.com/canton-foundation/canton-dev-fund/issues/187#issuecomment-4350403919_