GitHub user dosubot[bot] added a comment to the discussion: Superset → PostgreSQL: recurring SSL EOF and connection reset logs
<!-- Answer --> The recurring SSL EOF and connection reset logs on your PostgreSQL instance are most likely caused by how Superset manages its database connections and periodic background tasks, not by explicit health-check probes. Superset uses SQLAlchemy with NullPool by default, so each operation opens and closes a new connection, which can result in frequent connect/disconnect cycles and abrupt disconnects that PostgreSQL logs as SSL EOF or connection resets. Background tasks like cache warmup, report scheduling, and log pruning (run via Celery) also connect to the metadata database on a schedule, but these are not health checks and their frequency or behavior is not configurable via the Helm chart [see source](https://github.com/apache/superset/blob/829e4d92d91ceae4b43b1ed3b063ffe45377799c/superset/tasks/cache.py#L56-L331). Superset's Helm chart does not provide options to tune or disable database health checks, and Kubernetes liveness/readiness probes typically target Superset's HTTP endpoints, not the database directly [see source](https://github.com/apache/superset/issues/33604). If you have other infrastructure components or custom probes that connect to PostgreSQL, those could also contribute to the logs. To diagnose further, check if any other services or probes are connecting to your database, and consider increasing PostgreSQL logging to capture connection sources. If the logs are disruptive, you might try tuning connection pool settings, reviewing Celery schedules, or adjusting probe intervals elsewhere in your infrastructure. For reference on configuring Superset's database connection via Helm, see the [Helm chart values and connection logic](https://github.com/apache/superset/issues/33945). <!-- Dosu Comment Footer --> *To reply, just mention [@dosu](https://go.dosu.dev/dosubot).* --- <sup>How did I do? [Good](https://app.dosu.dev/response-feedback/6081d316-ea9b-4df6-90f1-6ed8393b8232?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/6081d316-ea9b-4df6-90f1-6ed8393b8232?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/6081d316-ea9b-4df6-90f1-6ed8393b8232?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/6081d316-ea9b-4df6-90f1-6ed8393b8232?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/6081d316-ea9b-4df6-90f1-6ed8393b8232?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/6081d316-ea9b-4df6-90f1-6ed8393b8232?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/6081d316-ea9b-4df6-90f1-6ed8393b8232?feedback_type=other)</sup> [](https://app.dosu.dev/a28d3c7e-a9d3-459e-9fb6-3a6f9ff4f357/ask?utm_source=github) [](https://cursor.com/link/prompt?text=The%20recurring%20SSL%20EOF%20and%20connection%20reset%20logs%20on%20your%20PostgreSQL%20instance%20are%20most%20likely%20caused%20by%20how%20Superset%20manages%20its%20database%20connections%20and%20periodic%20background%20tasks%2C%20not%20by%20explicit%20health-check%20probes.%20Superset%20uses%20SQLAlchemy%20with%20NullPool%20by%20default%2C%20so%20each%20operation%20opens%20and%20closes%20a%20new%20connection%2C%20which%20can%20result%20in%20frequent%20connect/disconnect%20cycles%20and%20abrupt%20disconnects%20that%20PostgreSQL%20logs%20as%20SSL%20EOF%20or%20connection%20resets.%20Background%20tasks%20like%20cache%20warmup%2C%20report%20scheduling%2C%20and%20log%20pruning%20%28run%20via%20Celery%29%20also%20connect%20to%20the%20metadata%20database %20on%20a%20schedule%2C%20but%20these%20are%20not%20health%20checks%20and%20their%20frequency%20or%20behavior%20is%20not%20configurable%20via%20the%20Helm%20chart%C2%A0%5Bsee%20source%5D%28https%3A//github.com/apache/superset/blob/829e4d92d91ceae4b43b1ed3b063ffe45377799c/superset/tasks/cache.py%23L56-L331%29.%0A%0ASuperset%27s%20Helm%20chart%20does%20not%20provide%20options%20to%20tune%20or%20disable%20database%20health%20checks%2C%20and%20Kubernetes%20liveness/readiness%20probes%20typically%20target%20Superset%27s%20HTTP%20endpoints%2C%20not%20the%20database%20directly%C2%A0%5Bsee%20source%5D%28https%3A//github.com/apache/superset/issues/33604%29.%20If%20you%20have%20other%20infrastructure%20components%20or%20custom%20probes%20that%20connect%20to%20PostgreSQL%2C%20those%20could%20also%20contribute%20to%20the%20logs.%0A%0ATo%20diagnose%20further%2C%20check%20if%20any%20other%20services%20or%20probes%20are%20connecting%20to%20your%20database%2C%20and%20consider%20increasing%20Postgre SQL%20logging%20to%20capture%20connection%20sources.%20If%20the%20logs%20are%20disruptive%2C%20you%20might%20try%20tuning%20connection%20pool%20settings%2C%20reviewing%20Celery%20schedules%2C%20or%20adjusting%20probe%20intervals%20elsewhere%20in%20your%20infrastructure.%0A%0AFor%20reference%20on%20configuring%20Superset%27s%20database%20connection%20via%20Helm%2C%20see%20the%20%5BHelm%20chart%20values%20and%20connection%20logic%5D%28https%3A//github.com/apache/superset/issues/33945%29.) [](https://go.dosu.dev/discord-bot) [](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/apache/superset/discussions/36420) GitHub link: https://github.com/apache/superset/discussions/36420#discussioncomment-15159311 ---- This is an automatically sent email for [email protected]. To unsubscribe, please send an email to: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
