Skip to content

Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.

ci(devcontainers): free runner disk on the default variant too (#1917)

Tejas Chopra904fa30586a8
diff --git a/.github/workflows/devcontainers.yml b/.github/workflows/devcontainers.yml
index 04552b676..3a90104d2 100644
--- a/.github/workflows/devcontainers.yml
+++ b/.github/workflows/devcontainers.yml
@@ -32,20 +32,17 @@ jobs:
steps:
- uses: actions/checkout@v6
- # The memory-stack devcontainer brings up Neo4j + Postgres +
- # Redis + Qdrant on top of the Docker base. PR #495 surfaced
- # that the GitHub-hosted runner now ships with too little
- # headroom for that stack: the runner's diagnostic log writer
- # hit "No space left on device" mid-smoke-test, killing the
- # job before `devcontainer exec` could finish. The default
- # devcontainer is light enough to pass on the same image.
+ # Both variants exhaust the GitHub runner's disk, so free space up front
+ # on every variant. memory-stack brings up Neo4j + Postgres + Redis +
+ # Qdrant (PR #495: the diagnostic log writer hit "No space left on device"
+ # mid-smoke-test). The default variant's post-create `uv sync` fills the
+ # disk installing the ML wheel set — it failed copying numpy into the venv
+ # with "No space left on device" (os error 28). Previously this ran only
+ # on memory-stack, which left the default variant with no cushion.
#
- # Reclaim ~14 GB by stripping preinstalled tools none of the
- # devcontainer paths use (Android SDK, .NET, Haskell). Run
- # ONLY on memory-stack — the default validate is fast and
- # benefits from the preinstalled toolcache.
- - name: Free runner disk (memory-stack only)
- if: matrix.name == 'memory-stack'
+ # Reclaim ~14 GB by stripping preinstalled tools none of the devcontainer
+ # paths use (Android SDK, .NET, Haskell).
+ - name: Free runner disk
uses: jlumbroso/[email protected]
with:
tool-cache: true