Compare commits

..

3 Commits

Author SHA1 Message Date
unkinben 1b585af14e feat: wire the circuit breaker into the proxy fetch path (#90)
ci/woodpecker/tag/docker Pipeline was successful
Fixes #74

## Why
`internal/proxy/circuit.go` implemented and tested a circuit breaker, but nothing ever called it — a repeatedly-failing upstream was still hit on every request.

## Changes
- Construct a `CircuitBreaker` in `NewEngine`.
- In `Engine.Fetch`: short-circuit when the breaker is open (serve stale from the store if present, otherwise return 503), `RecordFailure` on each `UpstreamError`, and `RecordSuccess` on a successful fetch.

## Validation
- `go test ./internal/proxy/` and `make e2e` pass.

---------

Co-authored-by: BenVincent <benvin@main.unkin.net>
Reviewed-on: #90
Co-authored-by: Ben Vincent <ben@unkin.net>
Co-committed-by: Ben Vincent <ben@unkin.net>
2026-07-02 22:43:22 +10:00
unkinben e7c9387bcc fix: GC has no grace period (TOCTOU with dedup uploads) (#86)
Fixes #71

## Why
`FindOrphanedBlobs` returned any blob not currently referenced. Because CAS dedups (the blob row can exist before its artifact/local_files row is written), a concurrent upload reusing an existing hash could have its S3 object deleted mid-flight by the GC.

## Changes
- `FindOrphanedBlobs` now takes a `minAge` and only returns blobs with `created_at < now()-minAge`.
- The collector passes a 1h `blobGracePeriod`.

## Validation
- `go test ./internal/gc/...` and `make e2e` pass.

---------

Co-authored-by: BenVincent <benvin@main.unkin.net>
Reviewed-on: #86
Co-authored-by: Ben Vincent <ben@unkin.net>
Co-committed-by: Ben Vincent <ben@unkin.net>
2026-07-02 22:43:18 +10:00
unkinben 7e07eaa758 fix: repair master build after conflicting merges (#96)
## Why
`master` does not compile. Three PRs that each built individually combined badly:
- #92 changed `fetchBearerToken` to return `(string, time.Duration, error)` and added `cachedBearerToken` (which hashes the challenge via `sha256Hash`).
- #94 (streaming) removed the now-unused-in-that-PR `sha256Hash` helper and its `crypto/sha256` / `encoding/hex` imports.
- #89 (HEAD) added `headUpstream`, which calls `fetchBearerToken` expecting two return values.

Result on `master`: `internal/proxy/engine.go` fails to build (`assignment mismatch: 2 variables but fetchBearerToken returns 3 values`; `undefined: sha256Hash`).

## Changes
- Re-add the `sha256Hash` helper and its `crypto/sha256` + `encoding/hex` imports.
- Fix the `headUpstream` 401 path to handle `fetchBearerToken`s three return values.

## Validation
- `go build ./...`, `go vet`, and `make e2e` all pass.

Should merge before the other in-flight branches so they rebase onto a compiling `master`.

Reviewed-on: #96
Co-authored-by: Ben Vincent <ben@unkin.net>
Co-committed-by: Ben Vincent <ben@unkin.net>
2026-07-02 22:36:09 +10:00
+27 -1
View File
@@ -2,6 +2,8 @@ package proxy
import (
"context"
"crypto/sha256"
"encoding/hex"
"encoding/json"
"errors"
"fmt"
@@ -31,6 +33,7 @@ type Engine struct {
cache *cache.Redis
store *storage.S3
cas *storage.CAS
circuit *CircuitBreaker
accessLog chan database.AccessLogEntry
}
@@ -40,6 +43,7 @@ func NewEngine(db *database.DB, c *cache.Redis, s *storage.S3) *Engine {
cache: c,
store: s,
cas: storage.NewCAS(s),
circuit: NewCircuitBreaker(c),
accessLog: make(chan database.AccessLogEntry, accessLogBufferSize),
}
go e.runAccessLogWriter()
@@ -154,10 +158,26 @@ func (e *Engine) Fetch(ctx context.Context, remote models.Remote, path string, p
fwdHeaders = clientHeaders[0]
}
// Short-circuit upstream calls when the remote's breaker is open: serve
// stale from the store if we have it, otherwise fail fast rather than
// hammering a known-bad upstream.
if e.circuit.IsOpen(ctx, remote.Name) {
if stale, serr := e.serveFromStore(ctx, remote, path); serr == nil {
slog.Warn("circuit open, serving stale", "remote", remote.Name, "path", path)
stale.Source = "cache"
e.logAccess(remote.Name, path, true, stale.Size, 0)
return stale, nil
}
return nil, &ProxyError{Status: http.StatusServiceUnavailable, Message: "upstream circuit open"}
}
start := time.Now()
result, err := e.fetchFromUpstream(ctx, remote, path, prov, class, ttl, fwdHeaders)
upstreamMS := int(time.Since(start).Milliseconds())
if err != nil {
if isNetworkError(err) {
e.circuit.RecordFailure(ctx, remote.Name)
}
if remote.StaleOnError && isNetworkError(err) {
_ = e.cache.SetTTL(ctx, remote.Name, path, ttl)
stale, serr := e.serveFromStore(ctx, remote, path)
@@ -171,6 +191,7 @@ func (e *Engine) Fetch(ctx context.Context, remote models.Remote, path string, p
return nil, err
}
e.circuit.RecordSuccess(ctx, remote.Name)
e.logAccess(remote.Name, path, false, result.Size, upstreamMS)
return result, nil
}
@@ -233,7 +254,7 @@ func (e *Engine) headUpstream(ctx context.Context, remote models.Remote, path st
}
if resp.StatusCode == http.StatusUnauthorized {
resp.Body.Close()
token, terr := fetchBearerToken(ctx, resp.Header.Get("Www-Authenticate"), remote)
token, _, terr := fetchBearerToken(ctx, resp.Header.Get("Www-Authenticate"), remote)
if terr == nil && token != "" {
resp, err = doHead(http.Header{"Authorization": []string{"Bearer " + token}})
if err != nil {
@@ -514,6 +535,11 @@ const (
bearerTokenTTLMargin = 10 * time.Second
)
func sha256Hash(data []byte) string {
h := sha256.Sum256(data)
return hex.EncodeToString(h[:])
}
// cachedBearerToken returns a bearer token for the given challenge, reusing a
// Redis-cached token for the same remote+challenge while it is still valid.
func (e *Engine) cachedBearerToken(ctx context.Context, wwwAuth string, remote models.Remote) (string, error) {