Skip to main content

Keep plan/apply logs forever

Feature Requests

We come from Terraform Cloud, and there any plan/apply will keep the logs forever. This is so useful when debugging something in the past.

However Spacelift only keeps logs for 2 months.

On top, even if it’s the latest run, the 2 months policy applies, so you could find yourself checking a stack without any logs…

Status: ❌ Rejected3 comments

Log in to comment and vote

Comments3

  • Jonah Kowall

    Team•

    Jul 12

    Thanks for the request. Run log retention defaults to 60 days but is customizable per account, so if you need longer retention, contact your account team: https://docs.spacelift.io/concepts/run. For permanent archives, webhooks and log streaming to stdout let you ship run data to your own storage as runs complete. “Forever” as a product default isn’t something we plan to offer, so I’m closing this, but the mechanisms above should cover compliance needs.

  • Black Breeze

    •

    Jun 5, 2025

    Thanks for this request! Before we dive into implementation, I'd love to understand the specific use case better.

    What problem are you trying to solve that requires accessing logs after 2+ months? Are you:
    - Debugging issues that surface much later?

    - Meeting compliance/audit requirements?

    - Analyzing patterns across historical runs?

    - Something else entirely?

    A few considerations on our end:

    • Cost impact: Forever storage significantly increases infrastructure costs for what appears to be edge case usage

    • Security: Long-term log retention in third-party systems can create data exposure risks

    • Existing solutions:

    - Self-hosted deployments let you control log retention periods

    - GraphQL API allows pulling logs programmatically for archival

    - SPACELIFT_RUN_LOGS_ON_STANDARD_OUTPUT_ENABLED lets you stream logs to your own systems

    What's not working with these current approaches for your specific need? Understanding your workflow would help us find the right solution - which might not require us storing logs forever at all.

    • Aquamarine Savanna

      •

      Sep 17, 2025

      Hey @Marcin Wyszynski, apologize for the late answer

      What problem are you trying to solve that requires accessing logs after 2+ months?

      One of the main problems is debugging issues that surface much later, like analysing plans/applies that could the root cause of an incident happening/discovered months later. However, analysing patterns across historical runs is also something we do.

      Cost impact: Forever storage significantly increases infrastructure costs for what appears to be edge case usage

      If this is the issue, we will be happy to provide our own AWS S3 bucket for you to store there the logs. If this means we can then set unlimited log retention and that those logs are going to be displayed for each run forever, we’re fine.

      I know there’s a current development in progress to allow to customise the log retention. Just in case it provides useful feedback, what we consider important is:

      • be able to set the log retention time to a custom value. If it’s per stack, the better.

      • be able to set a minimum number of runs to keep the logs for (per stack), independently on the log retention time. In this way, even if the log retention time has expired, there will always be runs with logs for each stack (the last N runs).