ENV VAR Clarification

[ryanflomenco]

EXECUTIONS_DATA_SAVE_ON_PROGRESS = save node execution data for errored and successful workflows? If this is set to false and save on error is set to true, what does the execution history for a failed workflow show?

EXECUTIONS_DATA_MAX_AGE = age in hours of executions before they are pruned. It says max, but that would imply executions will be maximum this age before pruning. What is the minimum age? If the execution data is pruned is there no remaining record of the execution in the database such that I can query it to determine the execution count?

EXECUTIONS_DATA_PRUNE_SOFT_DELETE_INTERVAL - What data is retained in a soft delete?

N8N_CONCURRENCY_PRODUCTION_LIMIT - does this replace the —concurrency command at runtime?

[n8n]

It looks like your topic is missing some important information. Could you provide the following if applicable.

• n8n version:
• Database (default: SQLite):
• n8n EXECUTIONS_PROCESS setting (default: own, main):
• Running n8n via (Docker, npm, n8n cloud, desktop app):
• Operating system:

[ria]

Hi @ryanflomenco

to answer your questions:

1. EXECUTIONS_DATA_SAVE_ON_PROGRESS = Saves data after each node execution. This means we update the same execution in the DB after every node finishes executing. When the flag is disabled we save only when the whole execution finishes, errors, or goes into waiting state.

2. EXECUTIONS_DATA_MAX_AGE = This variable operates together with EXECUTIONS_DATA_PRUNE_MAX_COUNT and basically sets a maximum number of executions to store in the database. Once you reach the limit count, n8n starts to delete the oldest execution records (you define what’s “old”). You can also turn off automatic pruning (though this is not recommended) Have a look here: Executions environment variables | n8n Docs

3. EXECUTIONS_DATA_PRUNE_SOFT_DELETE_INTERVAL = The soft deletion identifies all prunable executions and marks them for deletion (sets deleted timestamp) but still keeps data (including binary). This is quite an expensive query, hence the default interval is 60min.
   EXECUTIONS_DATA_PRUNE_HARD_DELETE_INTERVAL = The hard deletion then actually deletes the execution from DB and from disk. This is done in batches of 100, switching to 1s intervals until the total to prune is back down low enough. Once mostly caught up, hard deletion goes back to the set schedule, which is 15min by default.

4. N8N_CONCURRENCY_PRODUCTION_LIMIT = The --concurrency command is used in queue mode. See here - Concurrency control | n8n Docs

For reference:

Executions environment variables | n8n Docs

Hope this helps!