Hello All, Our environment consists of an indexer cluster scaled for 1 TB of data per day. On average, we have about 30 users logged running ad-hoc searches and about 40 scheduled searches running along side those queries. For 360 days of the year, the average CPU of our indexer cluster is no higher than 25%. But for 1 week of the year during the thanksgiving time range, we have about 65 users logged in, running ad-hoc queries, and loading multiple dashboards to monitor sales data during this time of the year. During this week of the year, the CPU on our indexers stays consistently at 90%-100% which we have attributed to many users loading dashboards with many panels simultaneously along with normal ad-hoc and scheduled searching. My question is, what recommendations are our there for combating this increased usage and prevent the CPU from being pegged at 100% for 1 week of the year? We are thinking about limiting the amount of searches each user is allowed to run concurrently but fear that many users will complain that their searches are queued. Any suggestions are much appreciated. Respectfully,
... View more