diff --git a/go.mod b/go.mod
index 5b748f52d..40d4b7872 100644
--- a/go.mod
+++ b/go.mod
@@ -6,7 +6,7 @@ require (
 	github.com/BurntSushi/toml v1.6.0
 	github.com/alecthomas/chroma/v2 v2.27.0
 	github.com/bradleyfalzon/ghinstallation/v2 v2.19.0
-	github.com/cloudbase/garm-provider-common v0.1.9
+	github.com/cloudbase/garm-provider-common v0.1.10-0.20260627162627-e00d7529cc6f
 	github.com/felixge/httpsnoop v1.1.0
 	github.com/gdamore/tcell/v2 v2.13.10
 	github.com/go-gormigrate/gormigrate/v2 v2.1.6
diff --git a/go.sum b/go.sum
index f1df2ca78..7b279994a 100644
--- a/go.sum
+++ b/go.sum
@@ -25,8 +25,8 @@ github.com/chzyer/test v1.0.0 h1:p3BQDXSxOhOG0P9z6/hGnII4LGiEPOYBhs8asl/fC04=
 github.com/chzyer/test v1.0.0/go.mod h1:2JlltgoNkt4TW/z9V/IzDdFaMTM2JPIi26O1pF38GC8=
 github.com/clipperhouse/uax29/v2 v2.7.0 h1:+gs4oBZ2gPfVrKPthwbMzWZDaAFPGYK72F0NJv2v7Vk=
 github.com/clipperhouse/uax29/v2 v2.7.0/go.mod h1:EFJ2TJMRUaplDxHKj1qAEhCtQPW2tJSwu5BF98AuoVM=
-github.com/cloudbase/garm-provider-common v0.1.9 h1:ZL53ma/j7BgMAqW/OJ/jCnx1MUd8hLcYxUTEFb/o/e0=
-github.com/cloudbase/garm-provider-common v0.1.9/go.mod h1:8tnJcLXtaMUDEUgX3MGLFEYnMpiCAKaKWPtzNnzEpAE=
+github.com/cloudbase/garm-provider-common v0.1.10-0.20260627162627-e00d7529cc6f h1:IalGGcSKBGUd+KFFn2OXO6SMcODyCkSO/SMXtm4nwy8=
+github.com/cloudbase/garm-provider-common v0.1.10-0.20260627162627-e00d7529cc6f/go.mod h1:i1KXJVzi5ouzbdu5BXjuA+rqk4nxIBVBEYrKReSERIU=
 github.com/cpuguy83/go-md2man/v2 v2.0.6/go.mod h1:oOW0eioCTA6cOiMLiUPZOpcVxMig6NIQQ7OS05n1F4g=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
diff --git a/runner/pool/pool.go b/runner/pool/pool.go
index b9414bef3..b575b60e5 100644
--- a/runner/pool/pool.go
+++ b/runner/pool/pool.go
@@ -1334,7 +1334,7 @@ func (r *basePoolManager) addRunnerToPool(pool params.Pool, aditionalLabels []st
 	}
 
 	if poolInstanceCount >= int64(pool.MaxRunners) {
-		return fmt.Errorf("max workers (%d) reached for pool %s", pool.MaxRunners, pool.ID)
+		return runnerErrors.NewNoCapacityError("max workers (%d) reached for pool %s", pool.MaxRunners, pool.ID)
 	}
 
 	if err := r.AddRunner(r.ctx, pool.ID, aditionalLabels); err != nil {
@@ -1412,8 +1412,6 @@ func (r *basePoolManager) retryFailedInstancesForOnePool(ctx context.Context, po
 
 	g, errCtx := errgroup.WithContext(ctx)
 	for _, instance := range existingInstances {
-		instance := instance
-
 		if instance.Status != commonParams.InstanceError {
 			continue
 		}
@@ -1472,7 +1470,10 @@ func (r *basePoolManager) retryFailedInstancesForOnePool(ctx context.Context, po
 				ctx, "queueing previously failed instance for retry",
 				"runner_name", instance.Name)
 			// Set instance to pending create and wait for retry.
-			if _, err := r.store.UpdateInstance(r.ctx, instance.Name, updateParams); err != nil {
+			// Use ForceUpdateInstance() here. It will ignore the instance transition from error to
+			// something other than a cleanup status (pending_delete/deleting). We don't really want to allow
+			// transitioning from error directly to "creating" otherwise.
+			if _, err := r.store.ForceUpdateInstance(r.ctx, instance.Name, updateParams); err != nil {
 				slog.With(slog.Any("error", err)).ErrorContext(
 					ctx, "failed to update runner status",
 					"runner_name", instance.Name)
@@ -2174,9 +2175,15 @@ func (r *basePoolManager) consumeQueuedJobs() error {
 				"pool_id", pool.ID,
 				"job_id", job.WorkflowJobID)
 			if err := r.addRunnerToPool(pool, jobLabels); err != nil {
-				slog.With(slog.Any("error", err)).ErrorContext(
-					r.ctx, "could not add runner to pool",
-					"pool_id", pool.ID)
+				if errors.Is(err, runnerErrors.ErrNoCapacity) {
+					slog.With(slog.Any("error", err)).InfoContext(
+						r.ctx, "could not add runner to pool",
+						"pool_id", pool.ID)
+				} else {
+					slog.With(slog.Any("error", err)).ErrorContext(
+						r.ctx, "could not add runner to pool",
+						"pool_id", pool.ID)
+				}
 				continue
 			}
 			slog.DebugContext(r.ctx, "a new runner was added as a response to queued job",
@@ -2187,7 +2194,7 @@ func (r *basePoolManager) consumeQueuedJobs() error {
 		}
 
 		if !runnerCreated {
-			slog.WarnContext(
+			slog.InfoContext(
 				r.ctx, "could not create a runner for job; unlocking",
 				"job_id", job.WorkflowJobID)
 			if err := r.store.UnlockJob(r.ctx, job.WorkflowJobID, r.ID()); err != nil {
diff --git a/vendor/github.com/cloudbase/garm-provider-common/errors/errors.go b/vendor/github.com/cloudbase/garm-provider-common/errors/errors.go
index 76e85d9cf..9a12c4ba7 100644
--- a/vendor/github.com/cloudbase/garm-provider-common/errors/errors.go
+++ b/vendor/github.com/cloudbase/garm-provider-common/errors/errors.go
@@ -32,6 +32,7 @@ var (
 	ErrTimeout          = NewTimeoutError("timed out")
 	ErrUnprocessable    = NewUnprocessableError("cannot process request")
 	ErrNoPoolsAvailable = NewNoPoolsAvailableError("no pools available")
+	ErrNoCapacity       = NewNoCapacityError("no capacity available")
 )
 
 type baseError struct {
@@ -43,7 +44,7 @@ func (b *baseError) Error() string {
 }
 
 // NewProviderError returns a new ProviderError
-func NewProviderError(msg string, a ...interface{}) error {
+func NewProviderError(msg string, a ...any) error {
 	return &ProviderError{
 		baseError{
 			msg: fmt.Sprintf(msg, a...),
@@ -66,7 +67,7 @@ func (p *ProviderError) Is(target error) bool {
 }
 
 // NewMissingSecretError returns a new MissingSecretError
-func NewMissingSecretError(msg string, a ...interface{}) error {
+func NewMissingSecretError(msg string, a ...any) error {
 	return &MissingSecretError{
 		baseError{
 			msg: fmt.Sprintf(msg, a...),
@@ -112,7 +113,7 @@ func (p *UnauthorizedError) Is(target error) bool {
 }
 
 // NewNotFoundError returns a new NotFoundError
-func NewNotFoundError(msg string, a ...interface{}) error {
+func NewNotFoundError(msg string, a ...any) error {
 	return &NotFoundError{
 		baseError{
 			msg: fmt.Sprintf(msg, a...),
@@ -158,7 +159,7 @@ func (p *DuplicateUserError) Is(target error) bool {
 }
 
 // NewBadRequestError returns a new BadRequestError
-func NewBadRequestError(msg string, a ...interface{}) error {
+func NewBadRequestError(msg string, a ...any) error {
 	return &BadRequestError{
 		baseError{
 			msg: fmt.Sprintf(msg, a...),
@@ -181,7 +182,7 @@ func (p *BadRequestError) Is(target error) bool {
 }
 
 // NewConflictError returns a new ConflictError
-func NewConflictError(msg string, a ...interface{}) error {
+func NewConflictError(msg string, a ...any) error {
 	return &ConflictError{
 		baseError{
 			msg: fmt.Sprintf(msg, a...),
@@ -204,7 +205,7 @@ func (p *ConflictError) Is(target error) bool {
 }
 
 // NewTimeoutError returns a new TimoutError
-func NewTimeoutError(msg string, a ...interface{}) error {
+func NewTimeoutError(msg string, a ...any) error {
 	return &TimoutError{
 		baseError{
 			msg: fmt.Sprintf(msg, a...),
@@ -227,15 +228,15 @@ func (p *TimoutError) Is(target error) bool {
 }
 
 // NewUnprocessableError returns a new UnprocessableError
-func NewUnprocessableError(msg string, a ...interface{}) error {
-	return &TimoutError{
+func NewUnprocessableError(msg string, a ...any) error {
+	return &UnprocessableError{
 		baseError{
 			msg: fmt.Sprintf(msg, a...),
 		},
 	}
 }
 
-// TimoutError is returned when an operation times out.
+// UnprocessableError is returned when a request cannot be processed.
 type UnprocessableError struct {
 	baseError
 }
@@ -249,16 +250,16 @@ func (p *UnprocessableError) Is(target error) bool {
 	return ok
 }
 
-// NewNoPoolsAvailableError returns a new UnprocessableError
-func NewNoPoolsAvailableError(msg string, a ...interface{}) error {
-	return &TimoutError{
+// NewNoPoolsAvailableError returns a new NoPoolsAvailableError
+func NewNoPoolsAvailableError(msg string, a ...any) error {
+	return &NoPoolsAvailableError{
 		baseError{
 			msg: fmt.Sprintf(msg, a...),
 		},
 	}
 }
 
-// NoPoolsAvailableError is returned when anthere are not pools available.
+// NoPoolsAvailableError is returned when there are no pools available.
 type NoPoolsAvailableError struct {
 	baseError
 }
@@ -271,3 +272,26 @@ func (p *NoPoolsAvailableError) Is(target error) bool {
 	_, ok := target.(*NoPoolsAvailableError)
 	return ok
 }
+
+// NewNoCapacityError returns a new NoCapacityError
+func NewNoCapacityError(msg string, a ...any) error {
+	return &NoCapacityError{
+		baseError{
+			msg: fmt.Sprintf(msg, a...),
+		},
+	}
+}
+
+// NoCapacityError is returned when there is no capacity available.
+type NoCapacityError struct {
+	baseError
+}
+
+func (p *NoCapacityError) Is(target error) bool {
+	if target == nil {
+		return false
+	}
+
+	_, ok := target.(*NoCapacityError)
+	return ok
+}
diff --git a/vendor/modules.txt b/vendor/modules.txt
index 21e24def6..bc027c676 100644
--- a/vendor/modules.txt
+++ b/vendor/modules.txt
@@ -25,7 +25,7 @@ github.com/chzyer/readline
 # github.com/clipperhouse/uax29/v2 v2.7.0
 ## explicit; go 1.18
 github.com/clipperhouse/uax29/v2/graphemes
-# github.com/cloudbase/garm-provider-common v0.1.9
+# github.com/cloudbase/garm-provider-common v0.1.10-0.20260627162627-e00d7529cc6f
 ## explicit; go 1.25.0
 github.com/cloudbase/garm-provider-common/cloudconfig
 github.com/cloudbase/garm-provider-common/defaults