- use issues
- Implement HEAD /v2/ WARN[0054] echo server HEAD:/v2/ status=405 latency=3.434µs host=localhost:8088 ip=127.0.0.1
- preload has confusing message "loaded 106 images to the file system cache " doesn't actually match the image count
- handler pullAndCache has a problem. The mem cache is updated outside of a "transaction" that pulls the image so - multiple pulls could overlap. upstream.Get should take an optional func pointer that updates the in-mem cache BEFORE the puller goroutine signals waiters.
- doneGet - refactor like imgpull
- Support upstream encoded into image url in case its not possible to configure containerd. E.g.:
image: in-cluster-mirror:8181/gcr.io/google-containers/echoserver:1.10
- requires doubling the API...
- Support multiple architectures for pre-loading
--arch=choice1,choice2
- Special handling for
latest
tag? - OTEL instrumentation
- number of images
- file system storage size
- memory size
- pulls over time
- remote IPs
- Base URL support (already in echo scaffolding)
- Support
--log-request-headers
- add CMD api to the openapi spec
- enable swagger UI