Repository navigation
Stabilize integration tests against auth-required, high-latency environments - #4700
Merged
Merged
Conversation
Fixes three groups of tests that fail nightly against Redis Enterprise: - RedisClientTest: checkPoolOverflow, invalidClientName and getNumActiveReturnsTheCorrectNumber built clients without credentials. Since the switch to RESP3 auto-negotiation (#4468) the handshake sends HELLO 3 eagerly, which auth-required servers reject with NOAUTH before any command runs. Use the endpoint's client config builder, which supplies credentials, like the sibling tests do. - ClientSideCacheFunctionalityTest.pendingInvalidationMessagesTest waits for oll=0/omem=0 in CLIENT LIST, which the Enterprise proxy does not faithfully expose. Disable on the 're' env, matching sibling tests. - HealthCheckIntegrationTest: probe strategies that open a real connection used 50-500 ms timeouts, below the RTT of remote endpoints. Raise them; strategies with synchronous lambdas are left untouched. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Test Resultsββ206 filesβ Β±0ββββ206 suitesβ Β±0βββ7m 52s β±οΈ +13s Results for commit 108bebd.βΒ± Comparison against base commit ecfed3d. β»οΈ This comment has been updated with latest results. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Fixes three groups of integration tests that fail nightly in the Redis Enterprise CI environment (auth-required endpoints, ~50β150 ms network RTT) while passing against the local OSS docker env:
RedisClientTestβcheckPoolOverflow,invalidClientNameandgetNumActiveReturnsTheCorrectNumberbuilt clients with a bareDefaultJedisClientConfig(no credentials). Since the switch to RESP3 auto-negotiation (Switch to RESP3 as default protocol versionΒ #4468), the handshake eagerly sendsHELLO 3, which auth-required servers reject withNOAUTHbefore any command runs (forinvalidClientName, before client-name validation, so the test caught the wrong exception). The tests now use the endpoint's client config builder, which supplies credentials β same as sibling tests.ClientSideCacheFunctionalityTest.pendingInvalidationMessagesTestβ waits foroll=0/omem=0inCLIENT LISTfor the tracked connection, which the Enterprise DMC proxy does not faithfully expose, so the awaitility condition times out. Disabled on thereenv via@ConditionalOnEnv, matching the sibling tests in the same class.HealthCheckIntegrationTestβ probe strategies that open a real connection (TCP + HELLO/AUTH + PING) used 50β500 ms timeouts, below the RTT of remote endpoints, so all probes failed and the health check resolved UNHEALTHY. Timeouts raised to 1000β2000 ms; strategies with synchronous lambdas (no network I/O) are left untouched.Why
The nightly "Jedis RE Integration Tests" workflow has been red for over a week; these three groups account for the deterministic failures (see e.g. run 32432112302). The remaining flaky failure (
advancedPubsubWithClientCacheClassCastException) is a product bug fixed separately.Testing
(timeout + delay) Γ probesin the adjusted health-check test stays within the existing 5 s latch.π€ Generated with Claude Code
Note
Low Risk
Test-only adjustments (credentials, env skip, timeouts). No production, auth, or pooling logic is changed.
Overview
Makes a handful of integration tests pass against Redis Enterprise (auth-required, higher RTT) without changing production code.
RedisClientTestnow buildscheckPoolOverflow,invalidClientName, andgetNumActiveReturnsTheCorrectNumberfrom the endpointβs client config so HELLO/AUTH credentials are present (bareDefaultJedisClientConfigwas failing withNOAUTHafter RESP3 auto-negotiation).pendingInvalidationMessagesTestis skipped on Enterprise because it depends on OSSCLIENT LISToll/omemfields the proxy does not expose.Health-check tests that open a real connection raise probe interval/timeout (to 1β2s) so remote RTT does not mark the endpoint UNHEALTHY; in-memory probe tests are unchanged.
Reviewed by Cursor Bugbot for commit 108bebd. Bugbot is set up for automated code reviews on this repo. Configure here.