Reduce CI load, remove unnecessary Namespace actions (tensorzero#3338)

* Reduce CI load

* Remove unnecessary Namespace actions

* Remove unnecessary Namespace actions

* Make building steps 2x8

* Make building steps 2x8

* Revert "Make building steps 2x8"

This reverts commit a103af7.

* Make test name obvious

* Adjust MQ live tests

* CH tests 16x32

---------

Co-authored-by: Gabriel Bianconi <[email protected]>
Co-authored-by: Aaron Hill <[email protected]>

Author: GabrielBianconi

.github/workflows/build-gateway-container.yml

@@ -5,7 +5,7 @@ on:
 
 jobs:
   build-gateway-container:
-    runs-on: namespace-profile-tensorzero-32x64;overrides.cache-tag=build-gateway-cache
+    runs-on: namespace-profile-tensorzero-2x8;overrides.cache-tag=build-gateway-cache
 
     steps:
       # TODO - investigate why using the Namespace checkout action causes

.github/workflows/build-ui-container.yml

@@ -5,7 +5,7 @@ on:
 
 jobs:
   build-ui-container:
-    runs-on: namespace-profile-tensorzero-32x64;overrides.cache-tag=build-ui-cache
+    runs-on: namespace-profile-tensorzero-2x8;overrides.cache-tag=build-ui-cache
 
     steps:
       - uses: namespacelabs/nscloud-checkout-action@953fed31a6113cc2347ca69c9d823743c65bc84b

.github/workflows/general.yml

@@ -179,17 +179,6 @@ jobs:
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
-      # We allow the namespace builder setup to fail on Dependabot PRs and PRs from forks
-      # (where the oidc token is not available)
-
-      - name: Install Namespace CLI
-        uses: namespacelabs/nscloud-setup@d1c625762f7c926a54bd39252efff0705fd11c64
-        continue-on-error: ${{ github.event.pull_request.head.repo.full_name != github.repository || github.actor == 'dependabot[bot]' }}
-
-      - name: Configure Namespace-powered Buildx
-        uses: namespacelabs/nscloud-setup-buildx-action@84ca8c58fdf372d6a4750476cd09b7b96ee778ca
-        continue-on-error: ${{ github.event.pull_request.head.repo.full_name != github.repository || github.actor == 'dependabot[bot]' }}
-
       # We deliberately install our MSRV here (rather than 'stable') to ensure that everything compiles with that version
       - name: Install Rust 1.86.0
         run: |
@@ -404,8 +393,7 @@ jobs:
     # We don't run many tests here, so use a normal runner with Github Actions caching
     # to avoid unnecessarily using Namespace credits (it should still always finish before
     # the main 'validate' job)
-    # runs-on: ${{ matrix.replicated && 'namespace-profile-tensorzero-32x64' || 'ubuntu-latest' }}
-    runs-on: namespace-profile-tensorzero-32x64
+    runs-on: ${{ matrix.replicated && 'namespace-profile-tensorzero-32x64' || 'namespace-profile-tensorzero-16x32' }}
     continue-on-error: ${{ matrix.clickhouse_version.allow_failure }}
     strategy:
       matrix:

.github/workflows/merge-queue.yml

@@ -57,7 +57,7 @@ jobs:
 
   live-tests:
     name: "live-tests (batch_writes: ${{ matrix.batch_writes }})"
-    runs-on: namespace-profile-tensorzero-32x64
+    runs-on: namespace-profile-tensorzero-16x32
     permissions:
       # Permission to checkout the repository
       contents: read
@@ -74,16 +74,13 @@ jobs:
       - name: Install gdb
         run: sudo apt-get update && sudo apt-get install -y gdb
 
-      - name: Warm up modal instances
+      - name: Warm up Modal instances
         run: |
           curl -H "Modal-Key: $MODAL_KEY" -H "Modal-Secret: $MODAL_SECRET" https://tensorzero--vllm-inference-vllm-inference.modal.run/docs > vllm_modal_logs.txt &
           curl -H "Modal-Key: $MODAL_KEY" -H "Modal-Secret: $MODAL_SECRET" https://tensorzero--sglang-inference-sglang-inference.modal.run/ > sglang_modal_logs.txt &
           # TODO: Re-enable once we can switch to a T4 GPU
           # curl -H "Modal-Key: $MODAL_KEY" -H "Modal-Secret: $MODAL_SECRET" https://tensorzero--vllm-gpt-oss-20b-serve.modal.run/ > vllm_gpt_oss_modal_logs.txt &
 
-      - name: Install Namespace CLI
-        uses: namespacelabs/nscloud-setup@d1c625762f7c926a54bd39252efff0705fd11c64
-
       - name: Cleanup disk space
         run: ./ci/free-disk-space.sh
 

.github/workflows/slash-command-regen-fixtures.yml

@@ -17,7 +17,7 @@ permissions:
 
 jobs:
   regen-fixtures:
-    runs-on: namespace-profile-tensorzero-32x64
+    runs-on: namespace-profile-tensorzero-8x16
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:

.github/workflows/ui-tests-e2e-model-inference-cache.yml

@@ -21,8 +21,8 @@ on:
         type: boolean
 
 jobs:
-  ui-tests-e2e:
-    runs-on: ${{ inputs.regen_cache && 'namespace-profile-tensorzero-32x64' || 'namespace-profile-tensorzero-16x32' }}
+  ui-tests-e2e-model-inference-cache:
+    runs-on: ${{ inputs.regen_cache && 'namespace-profile-tensorzero-8x16' || 'namespace-profile-tensorzero-16x32' }}
     # We currently only run this job when we have secrets available, since we need to use an S3 object_store
     # In the future, we might want to fix this so that it can run in PR CI for external (forked) PRs
     # For now, it just runs in the merge queue and on prs from the main repo