docs: improve timeout and retry sample #2630
Conversation
The timeout and retry sample used 60 seconds for both initial, max, and total timeout. It also used a 1.0 multiplier for the timeout value. This made it impossible to explain how to set an increasing RPC timeout value. It also rendered adding the DEADLINE_EXCEEDED error code as a retryable code superfluous.
product-auto-label
bot
added
size: xs
product-auto-label
bot
added
size: s
olavloite
added
the
owlbot:run
gcf-owl-bot
bot
removed
the
owlbot:run
| @@ -3419,6 +3419,48 @@ public void testRetryOnResourceExhausted() { | |||
| } | |||
| } | |||
|
|
|||
| @Test | |||
| public void testSampleRetrySettings() { | |||
There was a problem hiding this comment.
Can we carve it into a separate test class similar to what we do for most other samples code? It improves the discoverability of the test.
There was a problem hiding this comment.
We normally only do that for the integration tests that actually execute the sample code. See for example https://github.com/googleapis/java-spanner/blob/main/samples/snippets/src/test/java/com/example/spanner/DmlReturningSampleIT.java. This test does not run the actual sample code, as we don't have any infrastructure in place for running the samples against an in-mem mock server. Instead, it only verifies that not setting DEADLINE_EXCEEDED as one of the retryable codes will cause the operation to fail with DEADLINE_EXCEEDED.
There was a problem hiding this comment.
Nit: Consider adding the new test in a separate class. google-cloud-spanner/src/test/java/com/google/cloud/spanner/DatabaseClientImplTest.java is already overloaded for testing too many things.
There was a problem hiding this comment.
Moved the sample test to a separate class (and created an abstract base class for creating mock server tests, so it's easier to set up more of those test classes).
product-auto-label
bot
added
size: m
* docs: improve timeout and retry sample The timeout and retry sample used 60 seconds for both initial, max, and total timeout. It also used a 1.0 multiplier for the timeout value. This made it impossible to explain how to set an increasing RPC timeout value. It also rendered adding the DEADLINE_EXCEEDED error code as a retryable code superfluous. * 🦉 Updates from OwlBot post-processor See https://github.com/googleapis/repo-automation-bots/blob/main/packages/owl-bot/README.md * chore: update timeout values based on what is possible with other languages * test: move sample test to separate class --------- Co-authored-by: Owl Bot <gcf-owl-bot[bot]@users.noreply.github.com>
The timeout and retry sample used 60 seconds for both initial, max, and total timeout. It also used a 1.0 multiplier for the timeout value. This made it impossible to explain how to set an increasing RPC timeout value. It also rendered adding the DEADLINE_EXCEEDED error code as a retryable code superfluous.