fix: remove the working_directory attribute in the compactor config

This [flag](grafana/helm-charts@824e3da#diff-c09523f8c89a2fdb980fa9bb57eebcbb45a5950391a6bb8d44b549ea184388f6) was hardcoded before and was removed on the version 0.79.x of the chart. As a consequence, it seems that the attribute `compactor.working_directory' was taking its place and breaking our deployments. As a fix, I've decided to revert to the default configuration which uses the `/var/loki` directory (which is mounted as a persistent volume created automatically, as defined by our values).

aks/locals.tf

@@ -50,8 +50,7 @@ locals {
         }
         structuredConfig = {
           compactor = {
-            shared_store      = "azure"
-            working_directory = "/var/loki"
+            shared_store = "azure"
           }
         }
       })

eks/locals.tf

@@ -26,8 +26,7 @@ locals {
         }
         structuredConfig = {
           compactor = {
-            working_directory = "/data/compactor"
-            shared_store      = "s3"
+            shared_store = "s3"
           }
         }
       }

kind/locals.tf

@@ -30,8 +30,7 @@ locals {
         }
         structuredConfig = {
           compactor = {
-            working_directory = "/data/compactor"
-            shared_store      = "aws"
+            shared_store = "aws"
           }
         }
       }

sks/locals.tf

@@ -29,8 +29,7 @@ locals {
         }
         structuredConfig = {
           compactor = {
-            working_directory = "/data/compactor"
-            shared_store      = "s3"
+            shared_store = "s3"
           }
         }
       }