Add documentation, move concurrency near resources

values.yaml

@@ -35,11 +35,29 @@ mastodon:
     # Whether to perform DB migrations on `helm upgrade`.
     dbMigrate:
       enabled: true
+
+    # WARNING: deploySearch is potentially a very expensive job!
+    # Only enable this once at a time, when you deploy elasticsearch or when
+    # the upgrade notes for a new mastodon version request rebuilding search.
+    # Recommended use is via `-f mastodon.hooks.deploySearch.enabled=true`
+    # to ensure the job is only dispatched for a single upgrade when required.
+    # This job may take days to run on very large instances. Even small
+    # instances may take long enough to trigger helm's completion timeout, so
+    # DO NOT PANIC if helm complains; simply verify the job is still running.
+    #
+    # Builds or rebuilds the elasticsearch indices via `tootctl deploy search`
+    # with timing hooks to ensure the job runs immediately after install/upgrade
+    # and will be restarted if another, corrective upgrade is triggered.
+    # Please check the tootctl documentation and upgrade notes to pick values.
+    # 
+    # NOTE: The resource stanza set below is intentionally very conservative.
+    # Consider assigning a liberal chunk of your cluster's typical headroom.
     deploySearch:
       enabled: false
-      concurrency: 5
       resetChewy: true
-      only: "" # one index name. Possible values: instances, accounts, tags, statuses, public_statuses
+      # one index name. Possible values: instances, accounts, tags, statuses, public_statuses
+      only: ""
+      concurrency: 5
       resources: # this accepts any keys in a full container resources stanza.
         requests:
           cpu: 250m