Hey guys, I wanted to ask: Does the watchable schema feature need to send writes to db? We've enabled it with

--enable-experimental-watchable-schema-cache

and set

--datastore-schema-watch-heartbeat=300s

and saw a big spike in reads and writes to Postgres db (specially writes) which made it reach 100% cpu capacity. We have two separate workloads of spicedb in K8S: one that is read only and another that is read+write. We enabled the watchable schema on the read-only instance but not the read+write, is this a wrong approach? Thanks a lot 🙂 edit: spicedb version is v1.38.0, postgres 16.8, deployed to AKS: read-only has 3 replicas and read+write 2 replicas spicedb read-only args were:

- --grpc-enabled=true
          - --http-enabled=false
          - --dispatch-cluster-enabled=true
          - --metrics-enabled=true
          - --telemetry-endpoint=""
          - --datastore-engine=postgres
          - --dispatch-upstream-addr=kubernetes://spicedb.namespace:50053
          - --datastore-conn-pool-read-max-open=230
          - --datastore-conn-pool-read-min-open=115
          - --datastore-gc-window=1h
          - --datastore-gc-max-operation-time=3m
          - --log-level=warn
          - --enable-experimental-watchable-schema-cache=true
          - --datastore-schema-watch-heartbeat=300s