Internal error: Partial errors in view

I have two Couchbase Server nodes (v6.6.0 CE) in my cluster and one Sync Gateway (v2.8.2 CE) running on top of them. When querying a view through the admin API (port 4985) with stale=false, the Sync Gateway responds with:

{"error":"Internal Server Error","reason":"Internal error: Partial errors in view"}

I am clueless what to do at this point. I already tried to delete the view and recreate it with no change. What can I do to get to the root of this problem? Could it be the view code, the Couchbase or Sync Gateway version, my cluster setup or maybe a disk problem that leads to this error?

After searching through the logfiles, the final clue was
Failed to create cbsasl db update thread: Resource temporarily unavailable
which only occurred on one of the nodes. When checking the service via systemctl, this revealed:

● couchbase-server.service - Couchbase Server
   Loaded: loaded (/lib/systemd/system/couchbase-server.service; enabled; vendor preset: enabled)
   Active: active (running) since Wed 2021-04-28 11:12:57 CEST; 1 months 17 days ago
 Main PID: 22265 (beam.smp)
    Tasks: 4915 (limit: 4915)
   Memory: 5.7G

So the process unexpectedly hit the task limit after less than two months of operation. The other node only had 322 tasks. Restarting the service did the trick, but it’s still unclear how this happened.

To anyone stumbling across this problem in the future: Check your individual nodes for system resource problems and restart congested nodes. stale=false forces all nodes to update the view and any node having trouble doing so will lead to the error.