canton: plan DAR upgrade handling without node downtime

[Pessina]

Follow-up from PR #740.

Code:

mpc/chain-signatures/node/src/indexer_canton/mod.rs

Lines 164 to 173 in 93c8e00

	/// Canton JSON Ledger API configuration.
	///
	/// # Contract migration
	///
	/// When the Signer DAR is upgraded and redeployed, both `signer_contract_id`
	/// and `signer_template_id` change (similar to deploying a new Ethereum
	/// contract — the old address/ID is gone). Currently the MPC node requires a
	/// restart with updated CLI args to pick up the new IDs.
	///
	/// TODO: decide how to handle DAR upgrades without MPC node downtime.

Related check:

mpc/chain-signatures/node/src/indexer_canton/stream.rs

Lines 444 to 449 in 93c8e00

	// Check 3: ExercisedEvent with choice "SignBidirectional" on the pinned
	// Signer contract must exist in the same transaction. Exact contract ID
	// match since the operator pinned it via CLI.
	// NOTE: after a DAR upgrade/redeployment the contract ID changes — this
	// check will reject all events until the node is restarted with the new ID.
	// See CantonConfig migration TODO.

Context: when the Signer DAR is upgraded and redeployed, both signer_contract_id and signer_template_id change. Today the MPC node needs a restart with new CLI args, and the stream verification rejects events until the new contract ID is configured. Decide how Canton DAR upgrades should be handled without MPC node downtime.