Replay every unanimous upstream status, not just 4xx
ci/woodpecker/pr/test Pipeline was successful
ci/woodpecker/pr/build Pipeline was successful
ci/woodpecker/pr/pre-commit Pipeline was successful

openvoxdb does not reserve 5xx for its own faults: the same malformed
query is a 400 on /nodes and a 500 on /facts, and /metrics answers a flat
403, so a 4xx-only replay rule made pdbmux's behaviour depend on the
route. The meta and metrics handlers held their own copy of the gateway
error and bypassed the replay entirely.

- Replay any status from 400 up that every backend agreed on, with the
  backend's own body and content type.
- Keep 502 for backends disagreeing on the status, or a backend that
  answered nothing at all.
- Route /pdb/meta, /metrics and the pass-through path through the same
  rule as the merged query handlers.
- Count a unanimous 5xx as a failed round and let it fall back to a stale
  cache entry; only a unanimous 4xx stays exempt from both.
- Answer successful queries with openvoxdb's application/json;charset=utf-8.
This commit is contained in:
2026-09-13 13:35:21 +10:00
parent 5207a79ad4
commit 121bfacc2f
11 changed files with 541 additions and 126 deletions
+67 -1
View File
@@ -4,6 +4,7 @@ package main
import (
"context"
"net/url"
"strings"
"testing"
)
@@ -22,7 +23,7 @@ func TestRejectedQuerySurfacesUpstreamStatus(t *testing.T) {
t.Run(tc.name, func(t *testing.T) {
params := query(tc.q)
status, upstream := h.a.queryRaw(ctx, t, tc.path, params)
if !clientShaped(status) {
if !replayableStatus(status) || status >= 500 {
t.Fatalf("backend %s answered HTTP %d for %s, which is not the client error this test needs: %s",
h.a.name, status, tc.q, upstream)
}
@@ -56,3 +57,68 @@ func TestRejectedQuerySurfacesUpstreamStatus(t *testing.T) {
})
}
}
// openvoxdb answers the same malformed clause with 400 on /nodes and 500 on
// /facts — engine.clj's rewrite-fact-query runs an unguarded nth only for the
// facts entity. Both are the backends' real answer, so both have to reach the
// client rather than one of them becoming a 502.
func TestUnanimousServerErrorSurfacesUpstreamStatus(t *testing.T) {
ctx := context.Background()
params := query(`["=","name"]`)
status, upstream := h.a.queryRaw(ctx, t, factsPath, params)
if status < 500 {
t.Skipf("backend %s answered HTTP %d for the arity bug this test needs", h.a.name, status)
}
if other, _ := h.b.queryRaw(ctx, t, factsPath, params); other != status {
t.Fatalf("backends answered %d and %d, so the failure is not unanimous", status, other)
}
resp := rawGet(t, factsPath, params)
if resp.status != status {
t.Fatalf("pdbmux answered HTTP %d, want the upstream %d: %s", resp.status, status, resp.body)
}
if got, want := strings.TrimSpace(string(resp.body)), strings.TrimSpace(string(upstream)); want != "" && got != want {
t.Errorf("pdbmux body = %q, want openvoxdb's own explanation %q", got, want)
}
for _, b := range []*backend{h.a, h.b} {
if strings.Contains(string(resp.body), b.url) {
t.Errorf("replayed body names backend %s: %q", b.name, resp.body)
}
}
}
// /metrics and /pdb/meta had their own copy of the gateway error, so a status
// both backends agreed on never reached the client on those routes.
func TestUnanimousMetricsStatusIsReplayed(t *testing.T) {
ctx := context.Background()
for _, path := range []string{"/metrics/v1/mbeans", "/metrics/v2/list"} {
t.Run(path, func(t *testing.T) {
status, _ := h.a.queryRaw(ctx, t, path, nil)
if status < 400 {
t.Skipf("%s answers HTTP %d on this estate, so there is nothing to replay", path, status)
}
if other, _ := h.b.queryRaw(ctx, t, path, nil); other != status {
t.Fatalf("backends answered %d and %d, so the status is not unanimous", status, other)
}
if resp := rawGet(t, path, nil); resp.status != status {
t.Fatalf("pdbmux answered HTTP %d, want the upstream %d: %s", resp.status, status, resp.body)
}
})
}
}
// A client cannot tell pdbmux from a PuppetDB by the content type either.
func TestSuccessContentTypeMatchesUpstream(t *testing.T) {
ctx := context.Background()
params := url.Values{"limit": {"1"}}
_, _ = h.a.queryRaw(ctx, t, nodesPath, params)
resp := rawGet(t, nodesPath, params)
if resp.status != 200 {
t.Fatalf("status %d: %s", resp.status, resp.body)
}
if got := resp.header.Get("Content-Type"); got != "application/json;charset=utf-8" {
t.Errorf("Content-Type = %q, want openvoxdb's own %q", got, "application/json;charset=utf-8")
}
}