daemon: thread per-RPC op_id end-to-end

Today there's no way to correlate a CLI failure with a daemon log
line. operationLog records relative timing but no id, two concurrent
vm.start calls log indistinguishably, and the async
vmCreateOperationState.ID is user-facing yet never reaches the
journal. The root helper logs plain text to stderr while bangerd
logs JSON, so a merged journalctl is hard to grep across the
trust-boundary split.

Mint a per-RPC op id at dispatch entry, store it on context, and
include it as an "op_id" attr on every operationLog record. The
id is stamped onto every error response (including the early
short-circuit paths bad_version and unknown_method). rpc.Call
forwards the context op id on requests so a daemon RPC and the
helper RPCs it triggers all share one id. The helper now logs
JSON to match bangerd, adopts the inbound id, and emits a single
"helper rpc completed" / "helper rpc failed" line per call so
operators can see at a glance how long each privileged op took.

vmCreateOperationState.ID is now the same id dispatch generated
for vm.create.begin — one identifier between client status polls,
daemon logs, and helper logs.

The wire format gains two optional fields: rpc.Request.OpID and
rpc.ErrorResponse.OpID, both omitempty so older peers (and the
opposite direction) ignore them. ErrorResponse.Error() now appends
"(op-XXXXXX)" to its string form when set; existing callers that
just print err.Error() get the id for free.

Tests cover: dispatch stamps op_id on unknown_method, bad_version,
and handler-returned errors; rpc.Call exposes the typed
*ErrorResponse via errors.As so the CLI can read code/op_id; ctx
op_id is forwarded to the server in the request envelope.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Thales Maciel 2026-04-26 22:13:44 -03:00
parent b8c48765fb
commit e47b8146dc
No known key found for this signature in database
GPG key ID: 33112E6833C34679
16 changed files with 333 additions and 44 deletions

View file

@ -9,6 +9,7 @@ import (
"time"
"banger/internal/model"
"banger/internal/rpc"
)
func newDaemonLogger(w io.Writer, rawLevel string) (*slog.Logger, string, error) {
@ -35,9 +36,37 @@ func parseLogLevel(raw string) (slog.Level, string, error) {
}
}
func (d *Daemon) beginOperation(name string, attrs ...any) *operationLog {
// WithOpID stores the per-RPC correlation id on ctx. Re-exported
// from rpc so daemon-side call sites don't have to import rpc just
// for context plumbing. The dispatch layer calls this on every
// incoming request; capability hooks, lifecycle steps, and the
// privileged-ops shim that crosses into the root helper all read
// the id back via OpIDFromContext so a single id stitches the
// whole chain together in journalctl.
func WithOpID(ctx context.Context, opID string) context.Context {
return rpc.WithOpID(ctx, opID)
}
// OpIDFromContext returns the dispatch-assigned op id stored on
// ctx, or "" if none was set.
func OpIDFromContext(ctx context.Context) string {
return rpc.OpIDFromContext(ctx)
}
// beginOperation starts a logged operation. When ctx carries a
// dispatch-assigned op id (see WithOpID) every log line emitted
// through the returned operationLog includes it as an "op_id" attr,
// so the daemon journal can be greppable by id from the user's CLI
// error all the way down through capability hooks and the root
// helper.
func (d *Daemon) beginOperation(ctx context.Context, name string, attrs ...any) *operationLog {
opID := OpIDFromContext(ctx)
allAttrs := append([]any(nil), attrs...)
if opID != "" {
allAttrs = append([]any{"op_id", opID}, allAttrs...)
}
if d.logger != nil {
d.logger.Info("operation started", append([]any{"operation", name}, attrs...)...)
d.logger.Info("operation started", append([]any{"operation", name}, allAttrs...)...)
}
now := time.Now()
return &operationLog{
@ -45,7 +74,8 @@ func (d *Daemon) beginOperation(name string, attrs ...any) *operationLog {
name: name,
started: now,
last: now,
attrs: append([]any(nil), attrs...),
attrs: allAttrs,
opID: opID,
}
}
@ -55,6 +85,16 @@ type operationLog struct {
started time.Time
last time.Time
attrs []any
opID string
}
// OpID exposes the correlation id this operation was started with so
// dispatch can stamp it onto an outgoing error response.
func (o *operationLog) OpID() string {
if o == nil {
return ""
}
return o.opID
}
func (o *operationLog) stage(stage string, attrs ...any) {