<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Connectors on GCP Connectors for Apache Flink</title><link>https://flink-gcp.github.io/flink-connector-gcp/1.0/docs/connectors/</link><description>Recent content in Connectors on GCP Connectors for Apache Flink</description><generator>Hugo</generator><language>en-us</language><atom:link href="https://flink-gcp.github.io/flink-connector-gcp/1.0/docs/connectors/index.xml" rel="self" type="application/rss+xml"/><item><title>Delivery Guarantees</title><link>https://flink-gcp.github.io/flink-connector-gcp/1.0/docs/connectors/delivery-guarantees/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://flink-gcp.github.io/flink-connector-gcp/1.0/docs/connectors/delivery-guarantees/</guid><description>&lt;h1 id="delivery-guarantees"&gt;Delivery Guarantees&lt;a class="anchor" href="#delivery-guarantees"&gt;#&lt;/a&gt;&lt;/h1&gt;
&lt;p&gt;This page compares what a completed Flink checkpoint means for each sink and what can happen when
Flink restores an earlier checkpoint.&lt;/p&gt;
&lt;p&gt;The short version is that flushing before a checkpoint makes data durable, but does not by itself
prevent a restored job from writing the same record again.
Only BigQuery currently has checkpoint-coordinated exactly-once write methods.
Some other sinks can make a replay harmless when the record supplies a stable identity, but that is
not the same contract as a general-purpose exactly-once sink.&lt;/p&gt;</description></item></channel></rss>