RedisReactiveHealthIndicator should provide cluster details in cluster mode

[mhyeon-lee]

RedisReactiveHealthIndicator health is down with cluster mode.

Check different property if ReactiveRedisClusterConnection

private Health up(Health.Builder builder, Properties info) {
		return builder.up()
				.withDetail(RedisHealthIndicator.VERSION, info.getProperty(RedisHealthIndicator.REDIS_VERSION)).build();
	}

[wilkinsona]

@mp911de What's the equivalent of org.springframework.data.redis.connection.RedisClusterCommands.clusterGetClusterInfo() in the reactive API please?

[mp911de]

Redis Cluster commands (CLUSTER INFO) are not exposed through a reactive API. I filed DATAREDIS-1150 to provide a reactive variant of RedisClusterCommands.

[wilkinsona]

Thanks, @mp911de. Will that sneak into a Neumann SR or be a new feature in Ockham? If it's the latter, is there anything that we can do in the meantime?

[mp911de]

I need to discuss this with the team. Introducing new API in service releases is always subject to cause a certain degree of confusion. There's currently no workaround.

[marcus-bcl]

I've just hit this issue since upgrading to 2.3.0.RELEASE and have had to temporarily disable the Redis health endpoint, which isn't ideal (management.health.redis.enabled=false).

As a workaround until the change is implemented in spring-data-redis, would falling back to the non-reactive RedisHealthIndicator work? Something like:

@SpringBootApplication(exclude = RedisReactiveHealthContributorAutoConfiguration.class)

Appreciate any info you can provide - thanks!

[wilkinsona]

With thanks to @mp911de and the Data team, they've agreed to add the missing cluster-related APIs in a Spring Data Neumann service release. I believe it's currently slated for inclusion in Neumann-SR1 so, all being well, we should be able to fix this in Boot 2.3.1.

[andreacomo]

Waiting for Boot version 2.3.1, my workaround is:

1. disable default health check (management.health.redis.enabled=false)
2. Reimplement RedisReactiveHealthIndicator with a fixed version of up method like this:

private Health up(Health.Builder builder, Properties info, ReactiveRedisConnection connection) {
        if (connection instanceof ReactiveRedisClusterConnection) {
            List<Map<String, String>> details = getDetails(info);
            if (details.isEmpty()) {
                return builder.outOfService()
                        .build();
            } else {
                return builder.up()
                        .withDetail("nodes", details)
                        .build();
            }
        } else {
            return builder.up()
                    .withDetail("version", info.getProperty("redis_version"))
                    .build();
        }
    }

    private List<Map<String, String>> getDetails(Properties info) {
        return info.keySet().stream()
                        .map(String.class::cast)
                        .map(k -> k.substring(0, k.lastIndexOf(".")))
                        .distinct()
                        .sorted()
                        .map(node -> Map.of(
                                "node", node,
                                "redis_version", info.getProperty(node + ".redis_version"),
                                "role", info.getProperty(node + ".role"),
                                "uptime_in_days", info.getProperty(node + ".uptime_in_days")
                                )
                        )
                        .collect(Collectors.toList());
    }

Now my health check is UP.
Hope this can help

[christopher-nash]

We've experienced this same thing going from 2.2.7 to 2.2.8.

[wilkinsona]

@christopher-nash As far as I can tell, nothing has changed in this area between 2.2.7 and 2.2.8. If something was working for you in 2.2.7 and stopped working in 2.2.8, I don't think a fix for this issue will help. If you can reproduce the problem that you saw, please open a new issue with a minimal sample that works with 2.2.7 and fails with 2.2.8.