Overcoming dind image pull fails due to rate-limiting.

[braedongough]

Recently, we've encountered cases where a few of our runners will fail to pull the dind image due to rate limiting.

Failed to pull image "docker:dind": rpc error: code = Unknown desc = Error response from daemon: toomanyrequests: You have reached your pull rate limit. You may increase the limit by

I'm trying to evaluate the best way to overcome this, however, I'd love some clarifications on a few of the configuration options available.

1. If I set an imagePullSecret within the runnerdeployment will that apply to the image pull for the dind container? We have our own runner image hosted in ECR, so there haven't been any problems pulling the runner image. In my few simple tests, it appears that the imagePullSecret only applies to the actual runner container but please correct me if I'm wrong.

2. If I add a dockerRegistryMirror, will that apply also for pulling the dind image or is that only for the runner-image?

3. I've seen also this discussion whereby you add a docker-image argument in the to the actions-runner-controller deployment. In that solution, is it implied that would also be hosting your own dind image?

Apologies if these questions are trivial and obvious but any help with be appreciated!