CCPEDIAby Unity Nodes
ReleasesSplice 0.7.4

Splice 0.7.4

Digital AssetSplice

Upgrade Canton to 3.5.14. The enable-app-activity-record-and-traffic-ingestion and serve-app-activity-records-and-traffic configuration options have been removed. App activity records and sequencer traffic are now…

Runs on Canton 3.5.14

Open the release notes (opens in a new tab)docs.canton.network
  • Upgrade Canton to 3.5.14.
  • Scan app
    • The enable-app-activity-record-and-traffic-ingestion and serve-app-activity-records-and-traffic configuration options have been removed. App activity records and sequencer traffic are now always ingested, and app activity is always computed and served on the corresponding HTTP endpoints.
  • SV App
    • The SV app now exposes a splice.sv_vote_requests.active metric counting the active vote requests by their state relative to the SV (action_needed, in_progress, ready_to_close), allowing SV operators to alert on vote proposals that require their vote.
  • SV UI
    • The governance page now has a search field: paste a vote proposal contract ID to filter the Action Required, Inflight Votes, and Vote History sections down to that proposal, or press Enter to open its details page directly.
  • Scan & SV App
    • HTTP rate limiting has been extended with a global rate limiter applied across all operations, optional per-client-IP rate limiting (enabled by default at the global level), and an additional sustained rate limit enforced over a longer window on top of the existing per-second burst limit. The client IP is taken from the trusted, non-spoofable X-Envoy-External-Address header set by the Envoy/Istio ingress, falling back to the client-controlled X-Forwarded-For/ X-Real-Ip headers and finally the remote address only for requests that did not pass through the ingress. These can be tuned via the rate-limiting config keys.
When per-client-IP rate limiting is enabled, SV operators must ensure that the client IP used for rate limiting cannot be spoofed. Either configure rate-limiting.trusted-client-ip-header to a trusted, non-spoofable header set by your ingress/proxy (e.g. x-envoy-external-address for Istio deployments), or ensure that the X-Forwarded-For header contains the actual client IP as its first value and cannot be spoofed by clients. Otherwise, clients may bypass per-client-IP limits or cause other clients to be throttled by forging these headers.
The fallback to the client-controlled X-Forwarded-For/ X-Real-Ip headers can be disabled by setting rate-limiting.enable-client-provided-ip-headers to false. If no IP can be extracted no per IP rate limit is enforced. Comprehensive documentation about rate limiting configuration and infrastructure requirements (for SVs) will be added in the future.
  • Default rate limits have been adjusted.
  • CometBFT
    • Added a watchdog to restart cometbft when we detect that it is replaying messages. You must set watchdog.sequencerMetricsUrl: http://global-domain-SERIAL_ID-sequencer:10013/metrics and watchdog.mediatorMetricsUrl: http://global-domain-SERIAL_ID-mediator:10013/metrics in the cometbft helm values. If needed, the watchdog can be disabled through watchdog.enabled: false.
    • Bump the default deduplicationCacheSize to 1000000.
← All Splice releases