<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
    <id>https://www.nuntivus.com/blog/</id>
    <title>Nuntivus Blog</title>
    <updated>2026-04-15T00:00:00.000Z</updated>
    <generator>https://github.com/jpmonette/feed</generator>
    <link rel="alternate" href="https://www.nuntivus.com/blog/"/>
    <subtitle>Nuntivus Blog</subtitle>
    <icon>https://www.nuntivus.com/img/favicon.svg</icon>
    <entry>
        <title type="html"><![CDATA[Event delivery is more than an HTTP request]]></title>
        <id>https://www.nuntivus.com/blog/delivery-is-more-than-a-request/</id>
        <link href="https://www.nuntivus.com/blog/delivery-is-more-than-a-request/"/>
        <updated>2026-04-15T00:00:00.000Z</updated>
        <summary type="html"><![CDATA[Why customer-specific routing, retries and visibility belong in an explicit delivery layer.]]></summary>
        <content type="html"><![CDATA[<p>Sending a single request is straightforward. Delivering customer events repeatedly, with clear failure semantics and enough information to diagnose problems, is a different engineering problem.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="separate-acceptance-from-delivery">Separate acceptance from delivery<a href="https://www.nuntivus.com/blog/delivery-is-more-than-a-request/#separate-acceptance-from-delivery" class="hash-link" aria-label="Direct link to Separate acceptance from delivery" title="Direct link to Separate acceptance from delivery" translate="no">​</a></h2>
<p>An ingestion response and a successful downstream delivery are different facts. Nuntivus separates the API from asynchronous Worker processing, with PostgreSQL and NATS JetStream supporting its event flow. Operators can inspect what happened after acceptance instead of treating a network call as the entire lifecycle.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="make-contracts-visible">Make contracts visible<a href="https://www.nuntivus.com/blog/delivery-is-more-than-a-request/#make-contracts-visible" class="hash-link" aria-label="Direct link to Make contracts visible" title="Direct link to Make contracts visible" translate="no">​</a></h2>
<p>A customer contract captures event selection, delivery configuration and optional policies. Schema review, filtering and transformations make those choices explicit. Dry runs let operators inspect the pipeline without sending a delivery.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="design-for-duplicates">Design for duplicates<a href="https://www.nuntivus.com/blog/delivery-is-more-than-a-request/#design-for-duplicates" class="hash-link" aria-label="Direct link to Design for duplicates" title="Direct link to Design for duplicates" translate="no">​</a></h2>
<p>Timeouts create ambiguity: a remote service may have accepted a request even when its response was lost. Retrying cannot remove that ambiguity. Consumers should use stable event identity and their own idempotency strategy to prevent duplicate business effects.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="monitor-promises-not-assumptions">Monitor promises, not assumptions<a href="https://www.nuntivus.com/blog/delivery-is-more-than-a-request/#monitor-promises-not-assumptions" class="hash-link" aria-label="Direct link to Monitor promises, not assumptions" title="Direct link to Monitor promises, not assumptions" translate="no">​</a></h2>
<p>Success and latency SLOs describe observed delivery behavior. A monitoring target is not a switch that makes the network faster, and an empty traffic window is not proof of perfect reliability. Alerts and history help operators respond with context.</p>
<p>Read the <a class="" href="https://www.nuntivus.com/docs/reference/delivery-ordering-guarantees/">ordering guarantees</a>, <a class="" href="https://www.nuntivus.com/docs/reference/features/contract-slos/">contract SLOs</a> and <a class="" href="https://www.nuntivus.com/docs/reference/features/product-verification/">verification limitations</a> for the project's actual boundaries.</p>]]></content>
        <author>
            <name>Nuntivus contributors</name>
            <uri>https://github.com/nuntivus</uri>
        </author>
        <category label="architecture" term="architecture"/>
        <category label="reliability" term="reliability"/>
    </entry>
    <entry>
        <title type="html"><![CDATA[Introducing Nuntivus]]></title>
        <id>https://www.nuntivus.com/blog/introducing-nuntivus/</id>
        <link href="https://www.nuntivus.com/blog/introducing-nuntivus/"/>
        <updated>2026-04-15T00:00:00.000Z</updated>
        <summary type="html"><![CDATA[An open-source, self-hosted delivery layer between your business events and your customers.]]></summary>
        <content type="html"><![CDATA[<p>Every customer has a different integration story. One wants an HTTP webhook, another uses Kafka, and a third needs messages delivered to a queue. Application code can quickly become a collection of routing conditions, payload adapters and retry loops.</p>
<p>Nuntivus explores a different approach: a dedicated, self-hosted customer event delivery layer.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="publish-once-define-delivery-separately">Publish once, define delivery separately<a href="https://www.nuntivus.com/blog/introducing-nuntivus/#publish-once-define-delivery-separately" class="hash-link" aria-label="Direct link to Publish once, define delivery separately" title="Direct link to Publish once, define delivery separately" translate="no">​</a></h2>
<p>Sources bring events into the system. Customer contracts describe which events should be delivered and how to filter, validate and transform them. Shared destinations hold delivery configuration for HTTP, Kafka, RabbitMQ, NATS and SQS.</p>
<p>The administration dashboard brings the configuration and delivery history together, with connection tests, dry runs, retries, replays, SLOs and alerts.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="reliability-has-explicit-boundaries">Reliability has explicit boundaries<a href="https://www.nuntivus.com/blog/introducing-nuntivus/#reliability-has-explicit-boundaries" class="hash-link" aria-label="Direct link to Reliability has explicit boundaries" title="Direct link to Reliability has explicit boundaries" translate="no">​</a></h2>
<p>Nuntivus processes events at least once. Consumers must handle duplicate deliveries, and external side effects are not exactly once. Operational visibility is important, but it does not eliminate the need for backups, safe credential handling and downstream idempotency.</p>
<h2 class="anchor anchorTargetStickyNavbar_Vzrq" id="help-shape-the-project">Help shape the project<a href="https://www.nuntivus.com/blog/introducing-nuntivus/#help-shape-the-project" class="hash-link" aria-label="Direct link to Help shape the project" title="Direct link to Help shape the project" translate="no">​</a></h2>
<p>The project is in early development and is not yet production ready. APIs and schemas may change. Start with a local evaluation, review the documented limitations, and share feedback or issues on GitHub.</p>
<ul>
<li class=""><a class="" href="https://www.nuntivus.com/docs/quick-start/">Get started locally</a></li>
<li class=""><a class="" href="https://www.nuntivus.com/docs/reference/architecture/">Read the architecture</a></li>
<li class=""><a href="https://github.com/nuntivus/nuntivus-server" target="_blank" rel="noopener noreferrer" class="">Explore the repository</a></li>
</ul>]]></content>
        <author>
            <name>Nuntivus contributors</name>
            <uri>https://github.com/nuntivus</uri>
        </author>
        <category label="project" term="project"/>
        <category label="architecture" term="architecture"/>
    </entry>
</feed>