[FR]: Support custom clickhouse health check timeout

[ew0s]

Proposal summary

It'll be great to have ability to change duration of clickhouse (and other dependencies) healthcheck timeout through env vars

source

@Override
    protected Result check() {
        try {
            return template.nonTransaction(connection -> Mono.from(connection.createStatement("SELECT 1").execute())
                    .flatMap(result -> Mono.from(result.map((row, rowMetadata) -> row.get(0))))
                    .map(o -> Result.healthy()))
                    .block(Duration.ofSeconds({$ANALYTIC_DB_HEALTHCHECK_TIMEOUT})); // better to have ability to change this parameter through envs
        } catch (Exception ex) {
            return Result.unhealthy(ex);
        }
    }

Motivation

We have a restriction from db team to use only https connections to clickhouse, so that there are often cases when round trip in out infrastructure (network + ch healthcheck query) overlaps current timeout and k8s tend to restart pods in such cases.

Currently our backend pods restarted more than 200 times in 4 days.

[dsblank]

I investigated the feature request to make the ClickHouse health check timeout configurable. The issue was caused by a hardcoded 1-second timeout, which affected users with higher latency setups. I implemented a solution that allows users to set a custom timeout via the ANALYTICS_DB_HEALTH_CHECK_TIMEOUT_SECONDS environment variable. The changes are fully validated, and a pull request has been created for review. If any further issues arise, feel free to reach out for assistance.

✅ Solved the problem
✅ Pushed changes
✅ Created PR

Branch: ollie/issue-5397-custom-healthcheck-timeout-110903

---

---

_{This is an automated response. Responses may be inaccurate—please verify important details.
trace: 019c9a11-a715-7f5c-95aa-9c034f7161ce}