remove state management from keeper

the best keeper is a stateless keeper

Signed-off-by: Volkan Özçelik <[email protected]>

app/keeper/cmd/main.go

@@ -13,7 +13,6 @@ import (
 
 	"github.com/spiffe/spike/app/keeper/internal/env"
 	"github.com/spiffe/spike/app/keeper/internal/route/handle"
-	"github.com/spiffe/spike/app/keeper/internal/state"
 	"github.com/spiffe/spike/app/keeper/internal/trust"
 	"github.com/spiffe/spike/internal/auth"
 	"github.com/spiffe/spike/internal/config"
@@ -36,33 +35,6 @@ func main() {
 
 	trust.Authenticate(spiffeid)
 
-	keeperState := state.ReadAppState()
-
-	if keeperState == state.AppStateError {
-		log.FatalLn(
-			"SPIKE Keeper is in ERROR state. Manual intervention required.",
-		)
-	}
-
-	if keeperState == state.AppStateNotReady {
-		log.Log().Info(appName,
-			"msg", "SPIKE Keeper is not ready. Will send shards")
-
-		// go api.Contribute(source)
-		// go state.WaitForShards()
-	}
-
-	if keeperState == state.AppStateReady ||
-		keeperState == state.AppStateRecovering {
-		// TODO: implement this case
-		// 1. Transition to a RECOVERING state, if not done already
-		// 2. Contact peers to recompute shard.
-		// 3. Try forever.
-		// 4. If something is irrevocably irrecoverable transition to ERROR state.
-		// 5. When everything is back to normal, transition to READY state.
-		panic("I started, but I don't know what to do.")
-	}
-
 	log.Log().Info(appName, "msg", fmt.Sprintf("Started service: %s v%s",
 		appName, config.KeeperVersion))
 	if err := net.ServeWithPredicate(

jira.xml

@@ -14,6 +14,11 @@
     </target>
   </purpose>
   <next>
+    <issue>
+      we don't need state.ReadAppState()
+      and other state enums for keepers anymore
+      keepers are just dummy stateless keepers.
+    </issue>
     <issue>
       Invert shard generation flow.
     </issue>