gitlab-runner-helper out of memory during cache extraction

During cache extraction with the Kubernetes executor, gitlab-runner-helper runs out of memory:

Running with gitlab-runner 10.0.1 (e991d1b4)
  on kube-gce-europe-west2-staging-1024 (886c0a4c)
Using Kubernetes namespace: gitlab-runner
Using Kubernetes executor with image zerolab/drupal-ci:php ...
Waiting for pod gitlab-runner/runner-886c0a4c-project-64-concurrent-074m41 to be running, status is Pending
Waiting for pod gitlab-runner/runner-886c0a4c-project-64-concurrent-074m41 to be running, status is Pending
Waiting for pod gitlab-runner/runner-886c0a4c-project-64-concurrent-074m41 to be running, status is Pending
Running on runner-886c0a4c-project-64-concurrent-074m41 via kube-gce-europe-west2-staging-1024-3514937384-78kdx...
Cloning repository for master with git depth set to 3...
Cloning into '/builds/csp/csp8'...
Checking out 517ed785 as master...
Skipping Git submodules setup
Checking cache for default...
Downloading cache.zip from http://minio.gitlab-ci.svc:80/runner/runner/886c0a4c/project/64/default 
WARNING: web/vendor/behat/gherkin/src/Behat/Gherkin/Keywords/CachedArrayKeywords.php: write web/vendor/behat/gherkin/src/Behat/Gherkin/Keywords/CachedArrayKeywords.php: cannot allocate memory (suppressing repeats) 
/bin/bash: line 64:    51 Killed                  '/usr/bin/gitlab-runner-helper' "cache-extractor" "--file" "../../../csp/csp8/default/cache.zip" "--timeout" "10" "--url" "http://minio.gitlab-ci.svc:80/runner/runner/886c0a4c/project/64/default?<redacted>"
Failed to extract cache

The cache file is only 28MB (~60MB extracted) so I'm not sure what's going on here. The job container has a 1024MB memory limit, and the helper container has 256MB. The memory request is the same as the limit, so it's not overcommitted with other pods.