I’m experiencing an issue where my execution history is being aggressively pruned after approximately 13 hours, despite having configured a much longer retention period.
Environment:
- n8n version: 1.78.1
- Database: PostgreSQL
- Deployment: Docker container (2GB memory limit)
Configuration:
- EXECUTIONS_DATA_PRUNE=true
- EXECUTIONS_DATA_MAX_AGE=720 # 30 days
- EXECUTIONS_DATA_PRUNE_MAX_COUNT=0
Expected Behavior:
Execution history should be retained for 720 hours (30 days)
With EXECUTIONS_DATA_PRUNE_MAX_COUNT=0, there should be no count-based limitation
Actual Behavior:
Execution history is consistently truncated after ~13 hours
Latest query shows:
Oldest record: 2025-03-12 07:59:20
Newest record: 2025-03-12 21:16:30
Span: ~13 hours
Additional Details:
Memory usage is well within limits (74% of 2GB)
No relevant error messages in logs
Metrics collection is working correctly (visible in Grafana)
Database size is modest (1672 kB for execution table)
Has anyone else encountered this behavior? Is this a known issue? Any suggestions for troubleshooting would be appreciated.
I am unsure that:
EXECUTIONS_DATA_PRUNE_MAX_COUNT=0 is respected how you think. The documentation does not define 0 as an option for non limits. Try 50000 or something. I would also just turn EXECUTIONS_DATA_PRUNE=false off as well instead of to true.
Well I don’t post on forums until I’ve tried everything. So I’ve tried all numbers… 0, 10000000000, removing the directives completely… with and without EXECUTIONS_DATA_PRUNE… I’m sure I’ve tried every combination.
The =0 suggestion came from the forum:
In theory just turning off DATA_PRUNE and setting EXECUTIONS_DATA_PRUNE_MAX_COUNT to 0 would do the job but if you are using SQLite I would strongly recommend against doing this.
Where are these args set? Docker compose under n8n image?
Where was that suggestion? It could be true, especially if by a team member or admin, I haven’t used those args personally, I was just references the docs.
Gotcha.
Do you have logging enabled? Could you investigate the logs or provide some samples along that timeframe?
Hey @n8n_addict,
Please validate your YAML - it’s very sensitive to indentation.
Also, did you restart your instance? Any errors in the logs?
YAML is valid. No error logs but n8n has never outputted any logs anyway and I’ve given up trying.
I’ve been working on this for nearly 3 weeks, so at least 100 restarts and variations of docker-compose.yml files in that time.
It’s fine. I’ll just set up my own logging. ← accepted solution.
2 Likes