View, Retrieve, and Manage Logs
On this page
Ops Manager collects log information for both MongoDB processes and its agents. For MongoDB processes, you can access both real-time logs and on-disk logs.
The MongoDB logs provide the diagnostic logging information for your
mongod
andmongos
processes.The Agent logs provide insight into the behavior of your Ops Manager agents.
MongoDB Real-Time Logs
The MongoDB Agent issues the getLog
command with every
monitoring ping. This command collects log entries from RAM cache of
each MongoDB process.
Ops Manager enables real-time log collection by default. You can disable log collection for either all MongoDB deployments in an Ops Manager project or for individual MongoDB deployments. If you disable log collection, Ops Manager continues to display previously collected log entries.
View MongoDB Real-Time Logs
To access this feature, you must have privileges granted by one of the following roles:
Navigate to the Clusters view for your deployment.
If it is not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.
If it is not already displayed, select your desired project from the Projects menu in the navigation bar.
If it is not already displayed, click Deployment in the sidebar.
Click the Clusters view.
(Optional) For sharded clusters, filter which process type is listed.
The four buttons are listed in the following order, left to right: Shards, Configs, Mongos, and BIs.
Process | Displays |
---|---|
Shards | mongod processes that host your data. |
Configs | mongod processes that run as config
servers to store a sharded cluster's metadata. |
Mongos | mongos processes that route data in a sharded
cluster. |
BIs | BI processes that access data
in a sharded cluster. |
Enable or Disable Log Collection for a Deployment
Navigate to the Clusters view for your deployment.
If it is not already displayed, select the organization that contains your desired project from the Organizations menu in the navigation bar.
If it is not already displayed, select your desired project from the Projects menu in the navigation bar.
If it is not already displayed, click Deployment in the sidebar.
Click the Clusters view.
Enable or Disable Log Collection for the Project
MongoDB On-Disk Logs
Ops Manager collects on-disk logs even if the MongoDB instance is not
running. The MongoDB Agent collects the logs from the location you
specified in the MongoDB systemLog.path
configuration option. The
MongoDB on-disk logs are a subset of the real-time logs and therefore
less verbose.
You can configure log rotation for the on-disk logs. Ops Manager rotates logs by default.
This procedure rotates both system and audit logs for Ops Manager.
View MongoDB On-Disk Logs
To access this feature, you must have privileges granted by one of the following roles:
Click the deployment whose logs you want to download, then click Request Logs
Click the ellipsis ... icon on the line for any process, replica set, or sharded cluster in the project, then click Request Logs.
Select which types of log Ops Manager per process to collect and the maximum cumulative size of these logs.
To choose logs to download, perform the following actions:
Action | Purpose |
---|---|
Click MongoDB Logs | Gather logs from deployed MongoDB processes. This option isn't available for deployed MongoDB processes
if the systemLog.destination
property is set to |
Click FTDC Data | Gather the diagnostic data files from the (FTDC) collection mechanism,
such as server statistics and status messages. |
Click Automation Agent Logs | Gather logs from the deployed Automation Agents. |
Click Backup Agent Logs | Gather logs from all deployed Backup Agents. This differs from other logs. The logs collected are not limited to the selected hosts, but include all Backup Agent logs in the deployment. |
Click Monitoring Agent Logs | Gather logs from all deployed Monitoring Agent. This differs from other logs. The logs collected are not limited to the selected hosts, but include all Monitoring Agent logs in the deployment. |
Set Size per Log Type in MB | Enter the maximum cumulative uncompressed size in megabytes of all selected log files.
|
Example
You choose to collect 20 MB of logs from all processes in a replica set. This replica set has three mongod processes on two hosts:
host1:27017
host2:27017
host2:27018
Your deployment runs the following agents:
Automation Agent | host1 , host2 , host4 , host5 |
Backup Agent | host4 |
Monitoring Agent | host4 , host5 |
When you choose all the log types for this replica set, and limit to 20 MB per process, Ops Manager shows that the Estimated Total Size is 220 MB (11 processes * 20 MB) out of 20 GB.
Once the log collection begins, Ops Manager scans the log directories
for the mongod
processes and their associated FTDC from the
most current log entry until 20 MB of log files or the end of the
last log is collected. All the Monitoring and Backup Agents in the
deployment are also scanned.
The Backup Agent has 60 MB of logs.
Each MongoDB process (3) has 7 MB of logs plus 15 MB of FTDC data per process.
Each Monitoring Agent (2) has 30 MB of logs.
Each Automation Agent (2) has 12 MB of logs.
The total size of logs collected is 150 MB:
(20 + (3 * (7 + 15)) + (2 * 20) + (2 * 12)) = 150
The maximum of 20 MB of logs from the Backup Agent is collected.
All of the logs for each MongoDB process are collected: 7 MB of MongoDB + 15 MB of FTDC data.
The maximum of 20 MB of logs from each Monitoring Agent is collected.
All of the Automation Agent logs from
host1
andhost2
are collected.host4
andhost5
do not host any processes in the replica set.
The resulting archive structure within the downloaded archive is:
host1/27017/mongodb host1/27017/ftdc host1/automation_agent host2/27017/mongodb host2/27017/ftdc host2/27018/mongodb host2/27018/ftdc host2/automation_agent host4/backup_agent host4/monitoring_agent host5/monitoring_agent
(Optional) Redact sensitive information from the logs.
To anonymize your logs, select Replace IP addresses, hostnames, namespaces, strings with randomized values.
This option replaces IP addresses with a private range (192.168.x.x).
For hostnames, this option replaces only FQDN. Other hostnames
remain unchanged. Replacements follow a predictable pattern. For
example, if blue.strawberry
replaces one instance of the FQDN
test.internal
, blue.strawberry
replaces all other instances
of test.internal
as well.
Note
This does not use the $redact aggregation pipeline. That is a separate capability with a broader feature set.
View the progress on the Log Request History page.
Entry Status shows Collecting Logs... and automatically updates its status as log collection continues.
If Ops Manager fails to retrieve log files, click Retry to retrieve the failed log files again.
If a failure has occurred, you can still download the archive. Some of the requested log files will be missing.
Download the collected logs archive.
Click the Download icon.
The size listed for the archive on the Log Request History page is the uncompressed size. The archive consumes that amount of disk space on the target host once it is extracted.
This download can only be restarted and not resumed. If the download fails, you must download the logs again.
The archive is named mongodb-logfiles_<instance_or_process>_<ISO8601_Format_Date>.tar.gz
.
The extracted files use the following directory structure:
<host> automation_agent automation-agent-verbose.log automation-agent-verbose.log.<ISO8601_Format_Date> backup_agent backup-agent-verbose.log backup-agent-verbose.log.<ISO8601_Format_Date> monitoring_agent monitoring-agent-verbose.log monitoring-agent-verbose.log.<ISO8601_Format_Date> <replica_set> // Sharded Cluster Only <port> ftdc metrics.<ISO8601_Format_Date> metrics.interim mongodb mongodb.log mongodb.log.<ISO8601_Format_Date> <port> // Replica Set or Standalone ftdc metrics.<ISO8601_Format_Date> metrics.interim mongodb mongodb.log mongodb.log.<ISO8601_Format_Date>
Note
When you extract the tar
archive on a Microsoft Windows host,
use an archive extraction utility that supports PAX extended
headers.
Some Windows archive utilities have issues with PAX extended
headers for tar
.
Collected logs expire and are removed after 7 days. To extend the lifetime of a particular log file, click the extend link for that archive on the Log Request History page.
Configure Log Rotation
Ops Manager can rotate and compress logs for clusters that the MongoDB Agent manages. If the MongoDB Agent only monitors a cluster, it ignores that cluster's logs.
Important
If you're running MongoDB Enterprise version 5.0 or later and MongoDB Agent 11.0.13.7055 or later, you can:
Set separate rules for rotating server logs and audit logs.
Compress and delete audit logs using Ops Manager. For security reasons, we recommend managing your audit log compression and deletion outside of Ops Manager.
If you're running earlier versions of MongoDB Enterprise or the MongoDB Agent, Ops Manager:
Uses your System Log Rotation settings to rotate both the server logs and the audit logs.
Doesn't compress or delete audit logs. If you configure compression and deletion, Ops Manager applies these settings to the server logs only.
MongoDB Community users can rotate, compress, and delete the server logs only.
Note
When you use this feature, disable any platform-based log-rotation
services like logrotate
. Remove the reopen
and rename
flags from the process configuration files. If the MongoDB Agent only
monitors the cluster, that cluster may use platform-based services.
Enable log rotation.
Toggle System Log Rotation to ON to rotate server logs.
MongoDB Enterprise users running MongoDB Enterprise version 5.0 or later and MongoDB Agent 11.0.13.7055 and later can also toggle Audit Log Rotation to ON to rotate audit logs and configure audit log rotation separately.
If you're running earlier versions of MongoDB Enterprise or the MongoDB Agent, setting System Log Rotation to ON also rotates audit logs.
Set log rotation to OFF if you don't want Ops Manager to rotate its logs. Log rotation is OFF by default.
After you enable log rotation, Ops Manager displays additional log rotation settings.
Configure the log rotation settings.
Ops Manager rotates the logs on your MongoDB hosts per the following settings:
Field | Necessity | Action | Default |
---|---|---|---|
Size Threshold (MB) | Required | Ops Manager rotates log files that exceed this maximum log file
size. | 1000 |
Time Threshold (Hours) | Required | Ops Manager rotates logs that exceed this duration. | 24 |
Max Uncompressed Files | Optional | Log files can remain uncompressed until they exceed this number of files. Ops Manager compresses the oldest log files first. If you leave this setting empty, Ops Manager will use the default
of | 5 |
Max Percent of Disk | Optional | Log files can take up to this percent of disk space on your MongoDB host's log volume. Ops Manager deletes the oldest log files once they exceed this disk threshold. If you leave this setting empty, Ops Manager will use the default of
| 2% |
Total Number of Files | Optional | Total number of log files. If a number is not specified, the
total number of log files defaults to 0 and is determined
by other Rotate Logs settings. | 0 |
When you are done, click Save to review your changes.
Agent Logs
Ops Manager collects logs for all your MongoDB Agents.
View Agent Logs
To access this feature, you must have privileges granted by one of the following roles:
Filter the log entries.
To display logs for a different type of agent, use the View drop-down list.
To display logs for a specific hosts or MongoDB processes, click the gear icon and make your selections.
To clear filters, click the gear icon and click Remove Filters.
To download the selected logs, click the gear icon and click Download as CSV File.
Note
To view logs for a specific agent, you can alternatively click the Agents tab's All Agents list and then click view logs for the agent.
Configure Agent Log Rotation
If you use Automation to manage your cluster, follow this procedure to configure rotation of the Agent log files.
Note
If you haven't enabled Automation, see the following documentation for information about how to manually configure logging settings in the agent configuration files:
Edit the log settings.
Click the pencil icon to edit the Monitoring Agent or Backup Agent log settings:
Name | Type | Description | |
---|---|---|---|
Linux Log File Path | string | Conditional: Logs on a Linux host. The path to which the agent writes its logs on a Linux host. The suggested value is:
| |
Windows Log File Path | string | Conditional: Logs on a Windows host. The path to which the agent writes its logs on a Windows host. The suggested value is:
| |
Rotate Logs | Toggle | A toggle to select if the logs should be rotated. | |
Size Threshold (MB) | integer | The size where the logs rotate automatically. The default value
is 1000 . | |
Time Threshold (Hours) | integer | The duration of time when the logs rotate automatically. The
default value is 24 . | |
Max Uncompressed Files | integer | Optional. The greatest number of log files, including the
current log file, that should stay uncompressed. The suggested
value is 5 . | |
Max Percent of Disk | integer | Optional. The greatest percentage of disk space on your
MongoDB hosts that the logs should consume. The suggested
value is 2% . | |
Total Number of Files | integer | Optional. The total number of log files. If a number is not specified,
the total number of log files defaults to 0 and is determined by other
Rotate Logs settings. |
When you are done, click Save.
Ops Manager Logs
You can use Ops Manager to review a variety of log files:
You can change how long you keep some Ops Manager logs. Your company may need to keep log data for legal requirements. You can change your log retention policy to adhere to those requirements.