zylon-ai/private-gpt · critical · ValueError

Soft time limit should be less than or equal to hard time li

Error message

Soft time limit should be less than or equal to hard time limit

What it means

Config-validation ValueError from CelerySettings.validate_config: soft_time_limit must be <= hard_time_limit when both are set. Celery raises SoftTimeLimitExceeded inside the task at the soft limit so it can clean up, while the hard limit kills the worker process; an inversion of that ordering would make the soft signal useless, so the settings validator rejects it at startup.

Source

Thrown at private_gpt/settings/settings.py:1169

                else None
            )
        if "visibility_timeout" in data:
            data["visibility_timeout"] = (
                int(data["visibility_timeout"]) if data["visibility_timeout"] else None
            )
        super().__init__(**data)

    def validate_config(self) -> bool:
        # Check if visibility timeout is set when broker or backend is redis
        if self.broker_mode == "redis" or self.backend_mode == "redis":
            if not self.visibility_timeout:
                raise ValueError(
                    "Visibility timeout should be set when broker or backend is Redis"
                )

        if self.soft_time_limit:
            if self.hard_time_limit and self.soft_time_limit > self.hard_time_limit:
                raise ValueError(
                    "Soft time limit should be less than or equal to hard time limit"
                )

            if (
                self.visibility_timeout
                and self.visibility_timeout < self.soft_time_limit
            ):
                raise ValueError(
                    "Visibility timeout should be greater than soft time limit"
                )

        if self.hard_time_limit:
            if (
                self.visibility_timeout
                and self.visibility_timeout < self.hard_time_limit
            ):
                raise ValueError(
                    "Visibility timeout should be greater than hard time limit"

View on GitHub (pinned to 4a030776a3)

Solutions

  1. Make soft_time_limit <= hard_time_limit (typical pattern: soft = 80-90% of hard), e.g. soft 240 / hard 300.
  2. If the hard limit was lowered intentionally, lower the soft limit to match.
  3. Audit both environments' overrides (YAML + env vars) — the effective pair must satisfy the invariant everywhere.
  4. After editing, restart services so settings re-validate.

Example fix

# before
celery:
  soft_time_limit: 600
  hard_time_limit: 300
# after
celery:
  soft_time_limit: 240
  hard_time_limit: 300
Defensive patterns

Strategy: validation

Validate before calling

soft, hard = cfg['celery'].get('soft_time_limit'), cfg['celery'].get('hard_time_limit')
if soft and hard:
    assert soft <= hard, f'soft ({soft}) must be <= hard ({hard})'

Type guard

const limitsOrdered = (c: { soft_time_limit?: number; hard_time_limit?: number }): boolean =>
  c.soft_time_limit == null || c.hard_time_limit == null || c.soft_time_limit <= c.hard_time_limit;

Prevention

When it happens

Trigger: settings.yaml or env with soft_time_limit: 300 and hard_time_limit: 120; setting only one limit per environment (dev vs prod) and them drifting; unit confusion (seconds vs milliseconds) making soft appear larger.

Common situations: Ops tightening the hard limit in production without adjusting the soft limit; configs templated from examples with mismatched values; env var overrides applied to one limit only.

Related errors

AI-assisted analysis of zylon-ai/private-gpt@4a030776a3 (2026-08-15). Data as JSON: /api/errors/4fa9377591ded3bc. Report an issue: GitHub.