[feature] persist worker queues to db (#3042)

* persist queued worker tasks to database on shutdown, fill worker queues from database on startup

* ensure the tasks are sorted by creation time before pushing them

* add migration to insert WorkerTask{} into database, add test for worker task persistence

* add test for recovering worker queues from database

* quick tweak

* whoops we ended up with double cleaner job scheduling

* insert each task separately, because bun is throwing some reflection error??

* add specific checking of cancelled worker contexts

* add http request signing to deliveries recovered from database

* add test for outgoing public key ID being correctly set on delivery

* replace select with Queue.PopCtx()

* get rid of loop now we don't use it

* remove field now we don't use it

* ensure that signing func is set

* header values weren't being copied over 🤦

* use ptr for httpclient.Request in delivery

* move worker queue filling to later in server init process

* fix rebase issues

* make logging less shouty

* use slices.Delete() instead of copying / reslicing

* have database return tasks in ascending order instead of sorting them

* add a 1 minute timeout to persisting worker queues

internal/transport/delivery/worker.go

@@ -19,6 +19,7 @@ package delivery
 
 import (
 	"context"
+	"errors"
 	"slices"
 	"time"
 
@@ -160,6 +161,13 @@ func (w *Worker) process(ctx context.Context) bool {
 
 loop:
 	for {
+		// Before trying to get
+		// next delivery, check
+		// context still valid.
+		if ctx.Err() != nil {
+			return true
+		}
+
 		// Get next delivery.
 		dlv, ok := w.next(ctx)
 		if !ok {
@@ -195,16 +203,30 @@ loop:
 
 		// Attempt delivery of AP request.
 		rsp, retry, err := w.Client.DoOnce(
-			&dlv.Request,
+			dlv.Request,
 		)
 
-		if err == nil {
+		switch {
+		case err == nil:
 			// Ensure body closed.
 			_ = rsp.Body.Close()
 			continue loop
-		}
 
-		if !retry {
+		case errors.Is(err, context.Canceled) &&
+			ctx.Err() != nil:
+			// In the case of our own context
+			// being cancelled, push delivery
+			// back onto queue for persisting.
+			//
+			// Note we specifically check against
+			// context.Canceled here as it will
+			// be faster than the mutex lock of
+			// ctx.Err(), so gives an initial
+			// faster check in the if-clause.
+			w.Queue.Push(dlv)
+			continue loop
+
+		case !retry:
 			// Drop deliveries when no
 			// retry requested, or they
 			// reached max (either).
@@ -222,42 +244,36 @@ loop:
 
 // next gets the next available delivery, blocking until available if necessary.
 func (w *Worker) next(ctx context.Context) (*Delivery, bool) {
-loop:
-	for {
-		// Try pop next queued.
-		dlv, ok := w.Queue.Pop()
+	// Try a fast-pop of queued
+	// delivery before anything.
+	dlv, ok := w.Queue.Pop()
 
-		if !ok {
-			// Check the backlog.
-			if len(w.backlog) > 0 {
+	if !ok {
+		// Check the backlog.
+		if len(w.backlog) > 0 {
 
-				// Sort by 'next' time.
-				sortDeliveries(w.backlog)
+			// Sort by 'next' time.
+			sortDeliveries(w.backlog)
 
-				// Pop next delivery.
-				dlv := w.popBacklog()
+			// Pop next delivery.
+			dlv := w.popBacklog()
 
-				return dlv, true
-			}
-
-			select {
-			// Backlog is empty, we MUST
-			// block until next enqueued.
-			case <-w.Queue.Wait():
-				continue loop
-
-			// Worker was stopped.
-			case <-ctx.Done():
-				return nil, false
-			}
+			return dlv, true
 		}
 
-		// Replace request context for worker state canceling.
-		ctx := gtscontext.WithValues(ctx, dlv.Request.Context())
-		dlv.Request.Request = dlv.Request.Request.WithContext(ctx)
-
-		return dlv, true
+		// Block on next delivery push
+		// OR worker context canceled.
+		dlv, ok = w.Queue.PopCtx(ctx)
+		if !ok {
+			return nil, false
+		}
 	}
+
+	// Replace request context for worker state canceling.
+	ctx = gtscontext.WithValues(ctx, dlv.Request.Context())
+	dlv.Request.Request = dlv.Request.Request.WithContext(ctx)
+
+	return dlv, true
 }
 
 // popBacklog pops next available from the backlog.