Capabilities
Production Operations
Observability, resiliency, distributed coordination, and Relay Watch — the control plane for everything you shipped.
Production is where frameworks earn their keep: the 2 a.m. page, the slow dependency that takes everything with it, the job that must run on exactly one node.
What ships (observability): OpenTelemetry traces and metrics from a single
Nuvora.Nexus.Relay activity source, correlation ids end to end, and health
checks per subsystem — so “find the request” is a query, not an archaeology dig.
What ships (resiliency): retries with backoff, circuit breakers, rate and
concurrency limiters — as pipeline behaviors and consumer error policies, not
copy-pasted try/catch.
What ships (coordination): distributed locks, leader election, partitioning, and fenced leadership with cluster membership — “exactly one node does this” as a primitive, not a hope.
What ships (management): Relay Watch, the control plane. Services run a
lightweight agent (AddRelayManagementAgent); the host ingests registrations,
heartbeats and snapshots, derives KPIs, evaluates alert rules, and streams live
topology, saga, projection, DLQ and audit views to a React dashboard over
SignalR.
The stories behind it: The 2 A.M. Page Nobody Could Read, The Black Friday That Melted, and The Payout That Ran Twice.
Learn by building
The tutorials for this area, in order — each with a runnable sample.
021Observability: Telemetry, Metrics, Tracing, HealthYou cannot operate what you cannot see. Relay is instrumented with OpenTelemetry-native primitives out of the box — no wrappers, no proprietary format.
022Resiliency: Retries, Circuit Breakers, Rate & Concurrency LimitsDistributed systems fail in partial, transient ways: a database blips, a downstream API times out, a queue backs up.
023Distributed Coordination: Locks, Leader Election, PartitioningYou scale a service by running more copies of it — but now those copies can step on each other: two instances run the same nightly job, or both try to advance the same projection, or process the same key out of order.
025Reference Architecture: A Production Service End-to-EndThe previous twenty-four articles each isolated one capability.
030Fenced Leadership & Cluster MembershipArticle 023 elects a leader the simple way: a connection-bound ILeaderElector holds a PostgreSQL advisory lock for as long as its connection lives, and the database drops that lock when the process dies.
037Operational Event Queries: A Forensic Scan over the LogThe event store is an append-only log, and article 005 is emphatic about the rule that follows from that: never query the event store for reads.