The amount of Lively concurrent connections to OpenSearch Dashboards. If this number is constantly substantial, think about scaling your cluster.
The volume of requests on the _ppl API that failed as a result of a server difficulty or element limitation. Such as, a ask for may return HTTP position code 503 as a result of a VerificationException.
The number of queued responsibilities during the pressure merge thread pool. If your queue sizing is constantly high, think about scaling your cluster.
For each-node metric for the volume of graphs which have been evicted with the cache because of memory constraints or idle time. Specific evictions that take place as a consequence of index deletion are usually not counted. This metric is just applicable to approximate k-NN look for.
The amount of rejected duties inside the SQL lookup thread pool. If this quantity constantly grows, consider scaling your cluster.
With regards to the quantity of details and the scale in the cluster, upgrades may take anywhere from a couple of minutes to a few hrs to complete.
The amount of turned down jobs inside the look for thread pool. If this selection continuously grows, look at scaling your cluster.
The volume of queued jobs within the look for thread pool. In case the queue size is regularly higher, consider scaling your cluster. The most search queue measurement is one,000.
This metric will not be suitable and will be overlooked. The company will not use learn nodes as data nodes.
Automate the generation and administration of alerts & dashboards by using API, including the muting of alerts all through servicing durations
The utmost share of CPU resources utilized by the dedicated master nodes. We advocate increasing the dimensions of your instance type when this metric reaches sixty per cent.
Cluster configuration variations could possibly interrupt these functions before completion. We propose you use the /_tasks operation along Using these OpenSearch support operations to verify which the requests done productively.
The overall number of documents marked for deletion throughout all information nodes while in the cluster. These files no longer show up in search engine results, but OpenSearch only gets rid of deleted documents from disk all through phase merges. This metric boosts immediately after delete requests and decreases just after section merges.
The volume of follower indexes that were successfully made by a replication rule for a selected connection. AutoFollowNumFailedStartReplication