chore(ci): fix permanent instance selection condition

Due to continue-on-error directive use-permanent-instance step could not rely on failure() function.

.github/workflows/benchmark_gpu_core_crypto.yml

@@ -45,7 +45,7 @@ jobs:
       # This will allow to fallback on permanent instances running on Hyperstack.
       - name: Use permanent remote instance
         id: use-permanent-instance
-        if: ${{ env.SECRETS_AVAILABLE == 'true' && failure() }}
+        if: env.SECRETS_AVAILABLE == 'true' && steps.start-remote-instance.outcome == 'failure'
         run: |
           echo "runner_group=h100x1" >> "$GITHUB_OUTPUT"
 

.github/workflows/benchmark_gpu_erc20_common.yml

@@ -72,7 +72,9 @@ jobs:
       # This will allow to fallback on permanent instances running on Hyperstack.
       - name: Use permanent remote instance
         id: use-permanent-instance
-        if: ${{ env.SECRETS_AVAILABLE == 'true' && failure() && inputs.profile == 'single-h100' }}
+        if: env.SECRETS_AVAILABLE == 'true' &&
+          steps.start-remote-instance.outcome == 'failure' &&
+          inputs.profile == 'single-h100'
         run: |
           echo "runner_group=h100x1" >> "$GITHUB_OUTPUT"
 

.github/workflows/benchmark_gpu_integer_common.yml

@@ -136,7 +136,9 @@ jobs:
       # This will allow to fallback on permanent instances running on Hyperstack.
       - name: Use permanent remote instance
         id: use-permanent-instance
-        if: ${{ env.SECRETS_AVAILABLE == 'true' && failure() && inputs.profile == 'single-h100' }}
+        if: env.SECRETS_AVAILABLE == 'true' &&
+          steps.start-remote-instance.outcome == 'failure' &&
+          inputs.profile == 'single-h100'
         run: |
           echo "runner_group=h100x1" >> "$GITHUB_OUTPUT"
 

.github/workflows/gpu_fast_h100_tests.yml

@@ -91,7 +91,7 @@ jobs:
       # This will allow to fallback on permanent instances running on Hyperstack.
       - name: Use permanent remote instance
         id: use-permanent-instance
-        if: ${{ env.SECRETS_AVAILABLE == 'true' && steps.start-remote-instance.outcome == 'failure' }}
+        if: env.SECRETS_AVAILABLE == 'true' && steps.start-remote-instance.outcome == 'failure'
         run: |
           echo "runner_group=h100x1" >> "$GITHUB_OUTPUT"
 

.github/workflows/gpu_full_h100_tests.yml

@@ -42,7 +42,7 @@ jobs:
       # This will allow to fallback on permanent instances running on Hyperstack.
       - name: Use permanent remote instance
         id: use-permanent-instance
-        if: ${{ env.SECRETS_AVAILABLE == 'true' && failure() }}
+        if: env.SECRETS_AVAILABLE == 'true' && steps.start-remote-instance.outcome == 'failure'
         run: |
           echo "runner_group=h100x1" >> "$GITHUB_OUTPUT"
 

.github/workflows/gpu_signed_integer_h100_tests.yml

@@ -87,12 +87,12 @@ jobs:
           slab-url: ${{ secrets.SLAB_BASE_URL }}
           job-secret: ${{ secrets.JOB_SECRET }}
           backend: hyperstack
-          profile: single-h100
+          profile: single-h100-non-existing
 
       # This will allow to fallback on permanent instances running on Hyperstack.
       - name: Use permanent remote instance
         id: use-permanent-instance
-        if: ${{ env.SECRETS_AVAILABLE == 'true' && failure() }}
+        if: env.SECRETS_AVAILABLE == 'true' && steps.start-remote-instance.outcome == 'failure'
         run: |
           echo "runner_group=h100x1" >> "$GITHUB_OUTPUT"
 

.github/workflows/gpu_unsigned_integer_h100_tests.yml

@@ -91,7 +91,7 @@ jobs:
       # This will allow to fallback on permanent instances running on Hyperstack.
       - name: Use permanent remote instance
         id: use-permanent-instance
-        if: ${{ env.SECRETS_AVAILABLE == 'true' && failure() }}
+        if: env.SECRETS_AVAILABLE == 'true' && steps.start-remote-instance.outcome == 'failure'
         run: |
           echo "runner_group=h100x1" >> "$GITHUB_OUTPUT"