MongoDB Search triggers MongoDB Search alerts when the amount of CPU and memory used
by MongoDB Search processes reaches a specified threshold. If the search
process (mongot) runs out of memory, indexing and
queries fail. You can configure MongoDB Search alert conditions in the project
alert settings modal. You can also view MongoDB Search
metrics with cluster monitoring.
WARNING: If you shard a collection that already has a MongoDB Search index, you might experience a brief period of query downtime when the collection begins to appear on a shard. Also, if you add a shard for an already sharded collection that contains a MongoDB Search index, your search queries against that collection will fail until the initial sync process completes on the added shards. To learn more, see initial sync process.
Alert Conditions
You can configure the following alert conditions in the project-level alert settings page to trigger alerts.
Atlas Search: Index Replication Lag occurs if the approximate
number of milliseconds that MongoDB Search is behind in replicating changes from
the oplog of mongod is above or below the threshold.
Atlas Search: Index Size on Disk occurs if the total size of all
MongoDB Search indexes on disk in bytes is above or below the threshold.
Atlas Search: Max Number of Lucene Docs runs automatically by default.
This alert occurs if the upper bound number of Lucene docs used to store
MongoDB Search indexes for a given replica set or shard is above the threshold.
Atlas Search: Mongot stopped replication runs automatically by default.
This alert occurs only on dedicated Search Nodes if the replication is
interrupted by the MongoDB Search mongot process due to high disk utilization.
The pause replication threshold is 90% and the resume replication threshold
is 85% disk utilization.
The mongot process falls off the oplog if replication is paused for a
long time. Atlas rebuilds the index if the mongot process falls
off the oplog. However, you can prevent the mongot from falling off
the oplog if you upscale your search instances or delete MongoDB Search indexes.
Atlas Search: Number of Error Queries occurs if the number of
queries for which MongoDB Search is unable to return a response is above or
below the threshold.
Atlas Search: Number of Index Fields occurs if the total number
of unique fields present in the MongoDB Search index is above or below the
threshold.
Atlas Search: Number of Successful Queries occurs if the number
of queries for which MongoDB Search successfully returned a response is above or
below the threshold.
Atlas Search: Total Number of Queries occurs if the number of
queries submitted to MongoDB Search is above or below the threshold.
Atlas Search Opcounter: Delete occurs if the total number of
documents or fields (specified in the index definition) removed per
second is above or below the threshold.
Atlas Search Opcounter: Getmore occurs if the total number of
getmore commands run on all MongoDB Search queries per second is above or
below the threshold.
Atlas Search Opcounter: Insert occurs if the total number of
documents or fields (specified in the index definition) that MongoDB Search
indexes per second is above or below the threshold.
Atlas Search Opcounter: Update occurs if the total number of
documents or fields (specified in the index definition) that MongoDB Search
updates per second is above or below the threshold.
Insufficient disk space to support rebuilding search indexes runs
automatically by default. This alert occurs when your cluster
runs out of enough free disk space to support your MongoDB Search indexes.
Note
This alert might appear when Atlas automatically upgrades your search indexes to enable new features. Your cluster must have sufficient disk space for both the previous and new version of the index. If the rebuild increases disk space use to 90% or more, Atlas increases the cluster storage if you enabled auto-scaling. After the index upgrade completes, Atlas deletes the old version of the index, which frees up disk space.
Search Memory: Resident occurs
if the total bytes of resident memory occupied by the MongoDB Search process is
above or below the threshold.
Search Memory: Shared occurs if
the total bytes of shared memory occupied by the MongoDB Search process is above
or below the threshold.
Search Memory: Virtual occurs if
the total bytes of virtual memory occupied by the MongoDB Search process is
above or below the threshold.
Search Process: CPU (Kernel) % occurs if the percentage of time
the CPU spent servicing operating system calls for the MongoDB Search process is
above the threshold.
Search Process: CPU (User) %
occurs if the percentage of time the CPU spent servicing the MongoDB Search
process is above the threshold.
Search Process: Disk space used
occurs if the total bytes of disk space used by the MongoDB Search process is above the threshold.
Search Process: Ran out of memory
runs automatically by default. You can configure the alert setting
to disable this notification. This alert occurs when the search process (mongot)
runs out of memory. When the search process runs out of memory, indexing and queries fail.
Common Triggers
MongoDB Search alerts often occur when you try to build a large or complex search index. These indexes remain in the Initial Sync phase until you resolve the memory issue.
Fix the Immediate Problem
If the search process (mongot) runs out of memory or disk space,
you can upgrade your cluster to fix the
immediate problem. You can select a cluster tier with more memory,
storage, and IOPS.
Implement a Long-Term Solution
To prevent MongoDB Search alerts in the future, carefully review the Improve MongoDB Search Performance for MongoDB Search.
To optimize your indexes, we recommend deploying dedicated Search
nodes, which allows you to scale your
Atlas cluster and $search workloads independently.
Dedicated Search Nodes only run the mongot process and therefore
improve the availability, performance, and workload balancing of the
mongot process.
Monitor Your Progress
View the available MongoDB Search charts to monitor MongoDB Search metrics.
Monitor MongoDB Search metrics to evaluate and optimize your MongoDB Search indexes.
To learn more, see View Cluster Metrics