Alfresco Monitoring with Docker and ELK

Showing results for 
Search instead for 
Did you mean: 

Alfresco Monitoring with Docker and ELK

Active Member II
20 9 15.9K

1.     Project Objective

The aim of this blog is to show you how to create and run a Docker container with a full ELK (Elasticsearch, Logstash and Kibana) environment containing the necessary configuration and scripts to collect and present data to monitor your Alfresco application.

Elastic tools can ease the processing and manipulation of large amounts of data collected from logs, operating system, network, etc.

Elastic tools can be used to search for data such as errors, exceptions and debug entries and to present statistical information such as throughput and response times in a meaningful way. This information is very useful when monitoring and troubleshooting Alfresco systems.

2.     Install Docker on Host machine

Install Docker on your host machine (server) as per Docker website. Please note the Docker Community Edition is sufficient to run this project (

3.     Virtual Memory

Elasticsearch uses a hybrid mmapfs / niofs directory by default to store its indices. The default operating system limits on mmap counts is likely to be too low, which may result in out of memory exceptions.

On Linux, you can increase the limits by running the following command as root on the host machine:

# sysctl -w vm.max_map_count=262144

To set this value permanently, update the vm.max_map_count setting in /etc/sysctl.conf. To verify the value has been applied run:

# sysctl vm.max_map_count

4.     Download “Docker-ELK-Alfresco-Monitoring” container software

Download the software to create the Docker container from GitHub: and extract the files to the file system.

5.     Creating the Docker Image

Before creating the Docker image we need to configure access to Alfresco’s database from the Docker container. Assuming the files have been extracted to /opt/docker-projects/Docker-ELK-Alfresco-Monitoring-master, edit files and and set the access to the DB server as appropriate, for example:

#postgresql settings


db_url=jdbcSmiley Tongueostgresql://




Please make sure the database server allows for remote connections to Alfresco’s database. A couple of examples how to configure the database are shown here:

  • For MySQL

Access your database server as an administrator and grant the correct permissions i.e.

# mysql -u root -p

grant all privileges on alfresco.* to alfresco@'%' identified by 'admin';


The grant command is granting access to all tables in ‘alfresco’ database to ‘alfresco’ user from any host using ‘admin’ password.

Also make sure the bind-address parameter in my.cnf allows for external binding i.e. bind-address =


  • For PostgreSQL

Change the file ‘postgresql.conf’ to listen on all interfaces

listen_addresses = '*'

then add an entry in file ‘pg_hba.conf’ to allow connections from any host

host all all trust

Restart PostgreSQL database server to pick up the changes.

We have installed a small java application inside the container in /opt/activities folder that executes calls against the database configured in /opt/activities/ file.

For example to connect to PostgreSQL we have the following settings:


db_url=jdbcSmiley Tongueostgresql://




 We also need to set the timezone in the container, this can be done by editing the following entry in the script.

export TZ=GB


From the command line execute the following command to create the Docker image:

# docker build --tag=alfresco-elk /opt/docker-projects/Docker-Alfresco-ELK-Monitoring/

Sending build context to Docker daemon  188.9MB

Step 1/33 : FROM sebp/elk:530

530: Pulling from sebp/elk



6.     Creating the Docker Container

Once the Docker image has been created we can create the container from it by executing the following command:

# docker create -it -p 5601:5601 --name alfresco-elk alfresco-elk:latest


7.     Starting the Docker Container

Once the Docker container has been created it can be started with the following command:

# docker start alfresco-elk

Verify the ELK stack is running by accessing Kibana on http://localhost:5601 on the host machine.

At this point Elasticsearch and Kibana do not have any data…so we need to get Alfresco’s logstash agent up and running to feed some data to Elasticsearch.

8.     Starting logstash-agent

The logstash agent consists of logstash and some other scripts to capture entries from Alfresco log files, JVM stats using jstatbeat (, entries from Alfresco audit tables, DB slow queries, etc.

Copy the logstash-agent folder to a directory on all the servers running Alfresco or Solr applications.

Assuming you have copied logstash-agent folder to /opt/logstash-agent, edit the file /opt/alfresco-agent/ and set the following properties according to your own settings

export tomcatLogs=/opt/alfresco/tomcat/logs

export logstashAgentDir=/opt/logstash-agent

export logstashAgentLogs=${logstashAgentDir}/logs

export alfrescoELKServer=

 9.    Configuring Alfresco to generate data for monitoring

Alfresco needs some additional configuration to produce data to be sent to the monitoring Docker container.

9.1   Alfresco Logs

Alfresco logs i.e. alfresco.log, share.log, solr.log or the equivalent catalina.out can be parsed to provide information such as number of errors or exceptions over a period of time. We can also search these logs for specific data.


The first thing is to make sure the logs are displaying the full date time format at the beginning of each line. This is important so we can display the entries in the correct order.

Make sure in your log4j properties files (there is more than one) the file layout pattern is as follows:

log4j.appender.File.layout.ConversionPattern=%d{yyyy-MM-dd} %d{ABSOLUTE} %-5p [%c] [%t] %m%n

This will produce log entries with the date at the beginning of the line as this one:

2016-09-12 12:16:28,460 INFO  [org.alfresco.repo.admin] [localhost-startStop-1] Connected to database PostgreSQL version 9.3.6

Important Note: If you upload catalina files then don’t upload alfresco (alfresco, share, solr) log files for the same time period since they contain the same entries and you will end up with duplicate entries in the Log Analyser tool.


Once the logs are processed the resulting data is shown:

  • Number errors, warnings, debug, fatal messages, etc over time
  • Total number of errors, warnings, debug, fatal messages, etc
  • Common messages that may reflect issues with the application
  • Number of entries grouped by java class
  • Number of exceptions logged
  • All log files are searchable using ES (Elasticsearch) search syntax



9.2    Document Transformations

Alfresco performs document transformations for document previews, thumbnails, indexing content, etc. To monitor document transformations enable logging for class “TransformerLog”  by adding the following line to tomcat/shared/classes/alfresco/extension/ on all alfresco nodes:

The following is a sample output from alfresco.log file showing document transformation times, document extensions, transformer used, etc.

2016-07-14 18:24:56,003  DEBUG [content.transform.TransformerLog] [pool-14-thread-1] 0 xlsx png  INFO Calculate_Memory_Solr Beta 0.2.xlsx 200.6 KB 897 ms complex.JodConverter.Image<<Complex>>


Once Alfresco logs are processed the following data is shown for transformations:

  • Response time of transformation requests over time
  • Transformation throughput
  • Total count of transformations grouped by file type
  • Document size, transformation time, transformer used, etc



9.3    Tomcat Access Logs

Tomcat access logs can be used to monitor HTTP requests, throughput and response times. In order to get the right data format in the logs we need to add/replace the “Valve” entry in tomcat/conf/server.xml file, normally located at the end of the file, with this one below.




  prefix="access-" suffix=".log"

  pattern='%a %l %u %t "%r" %s %b "%{Referer}i" "%{User-agent}i" %D "%I"'




 For further clarification on the log pattern refer to:

Sample output from tomcat access log under tomcat/logs directory. The important fields here are the HTTP request, the HTTP response status i.e. 200 and the time taken to process the request i.e. 33 milliseconds - CN=Alfresco Repository Client, OU=Unknown, O=Alfresco Software Ltd., L=Maidenhead, ST=UK, C=GB [14/Jul/2016:18:49:45 +0100] "POST /alfresco/service/api/solr/modelsdiff HTTP/1.1" 200 37 "-" "Spring Surf via Apache HttpClient/3.1" 33 "http-bio-8443-exec-10"

Once the Tomcat access logs are processed the following data is shown: 

  • Response time of HTTP requests over time
  • HTTP traffic throughput
  • Total count of responses grouped by HTTP response code
  • Tomcat access logs files are searchable using ES (Elasticsearch) search syntax



9.4    Solr Searches

We can monitor Solr queries and response times by enabling debug for class SolrQueryHTTPClient by adding the following entry to tomcat/shared/classes/alfresco/extension/ on all Alfresco (front end) nodes:

Sample output from alfresco.log file showing Solr searches response times:

DEBUG [impl.solr.SolrQueryHTTPClient] [http-apr-8080-exec-6]    with: {"queryConsistency":"DEFAULT","textAttributes":[],"allAttributes":[],"templates":[{"template":"%(cm:name cm:title cm:description ia:whatEvent ia:descriptionEvent lnk:title lnk:description TEXT TAG)","name":"keywords"}],"authorities":["GROUP_EVERYONE","ROLE_ADMINISTRATOR","ROLE_AUTHENTICATED","admin"],"tenants":[""],"query":"((test.txt  AND (+TYPE:\"cm:content\" +TYPE:\"cm:folder\")) AND -TYPE:\"cm:thumbnail\" AND -TYPE:\"cm:failedThumbnail\" AND -TYPE:\"cm:rating\") AND NOT ASPECT:\"sys:hidden\"","locales":["en"],"defaultNamespace":"","defaultFTSFieldOperator":"OR","defaultFTSOperator":"OR"}

 2016-03-19 19:55:54,106 


DEBUG [impl.solr.SolrQueryHTTPClient] [http-apr-8080-exec-6] Got: 1 in 21 ms

Note: There is no specific transaction id to correlate the Solr search to the corresponding response. The best way to do this is to look at the time when the search and response were logged together with the java thread name, this should give you a match for the query and its response. 

Once Alfresco logs are processed the following data is shown for Solr searches:

  • Response time for Solr searches over time
  • Solr searches throughput
  • Solr queries, number of results found and individual response times



9.5    Database Monitoring

Database performance can be monitored with two different tools: p6spy and packetbeats. The main difference between these tools is that p6spy acts as a proxy jdbc driver and packetbeat is a network traffic sniffer. Also packetbeat can only sniff traffic for MySQL and PostgreSQL databases while p6spy can also do Oracle among others.



P6spy software is delivered as a jar file that needs to be placed in the application class path i.e. tomcat/lib/ folder. There are 3 steps to get p6spy configured and running.


  • Place p6spy jar file in tomcat/lib/ folder
  • Create file also in tomcat/lib/folder with the following configuration




dateformat=MM-dd-yy HH:mm:ssSmiley FrustratedS






# Update driver list correct driver i.e.

# driverlist=oracle.jdbc.OracleDriver

# driverlist=org.mariadb.jdbc.Driver

# driverlist=org.postgresql.Driver



# Location where spy.log file will be created



# Set the execution threshold to log queries taking longer than 1000 milliseconds (slow queries only)



Note: if there are no queries taking longer than the value in executionThreshod (in milliseconds) then the file will not be created.

Note: set the “logfile” variable to the logs folder inside the logstash-agent path as shown above.


  • Add entry to tomcat/conf/Catalina/localhost/alfresco.xml file


Example for PostgreSQL:








  url="jdbcSmiley Tongue6spySmiley Tongueostgresql://localhost/p6spy:alfresco"







Example for Oracle:








  url="jdbcSmiley Tongue6spySmiley Surprisedracle:thin:@"   







 Example for MariaDB:








  url="jdbcSmiley Tongue6spy:mariadb://localhost:3306/alfresco"






Once the spy.log file has been processed the following information is show:

  • DB Statements execution time over time
  • DB Statements throughput over time
  • Table showing individual DB statements and execution times
  • DB execution times by connection id 



9.6    Alfresco Auditing

If you want to audit Alfresco access you can enable auditing by adding the following entries to file:

# Enable auditing


Now you can monitor all the events generated by alfresco-access audit group.


Note: Only one of the logstash agents should collect Alfresco's audit data since the script gathers data for the whole cluster/solution. So edit the file logstash_agent/ in one of the other Alfresco nodes and set the variable collectAuditData to "yes" as indicated below:


Note: Also make sure you update the login credentials for Alfresco in the audit*sh files. Defaults to admin/admin.


10.    Starting and Stopping the logstash agent

The logstash agent script can be started from the command line with "./ start" as shown below:

./ start
Starting logstash
Starting jstatbeat
Starting dstat
Staring audit access script

and can be stopped with the command "./ stop" as shown below:

./ stop
Stopping logstash
Stopping jstatbeat
Stopping dstat
Stopping audit access script

11. Accessing the Dashboard

Finally access the dashboard by going to this URL http://<docker host IP>:5601 (use the IP of the server where you installed the Docker container) and clicking on the “Dashboard” link on the left panel and then click on the “Activities” link.



The data should be available for the selected time period.

Navigate to the other dashboards by clicking on the appropriate link.

11.    Accessing the container

To enter the running container use the following command:

# docker exec -i -t alfresco-elk bash

And to exit the container just type “exit” and you will find yourself back on the host machine.

12.    Stopping the container

To stop the container from running type the following command on the host machine:

Header 1

# docker stop alfresco-elk

13.    Removing the Docker Container

To delete the container you first need to stop the container and then run the following command:

# docker rm alfresco-elk

14.    Removing the Docker Image

To delete the container you first need to stop the container and then run the following command:

# docker rmi alfresco-elk:latest

15.   Firewall ports

If you have a firewall make sure the following ports are ope:

Redis: 6379

Kibana: 5601

Database server: this depends on the DB server being used i.e. PostgreSQL is 5432, MySQL 3306, etc

Happy Monitoring!!!


A big THANK YOU for this article.

I'll try ASAP!


Hi Miguel:

First of all, thanks for this monitoring setup project. It is very useful and instructive, and it provides nice insights of your Alfresco setup and config. I have three questions about it:

  • What is the minimum or recommended setup to be considered in a production environment ? Are audit, P6Spy DB or workflow & activies queries valid for production ?
  • I can see a bunch Info about GC, but what about some dashboard about JVM and threads in the way of Support Tools do ?
  • Would you obtain P6SPY logs if not finding slow queries (executionThreshold=1000) ?

Regarding the tutorial, I saw some minor point that I found during my setup:

  • 5. In creating Docker image: Edit my.cnf or similar for allowing remote connections to DB (bind-address =
  • 5. In workflows queries inside I had to put "act_hi_taskinst" table in CAPS (it may depend on your DB config)
  • 7. Starting the docker container: Verify the ELK stack is running by accessing Kibana in (not localhost)
  • 9.5. P6SPY: p6spy jar must be downloaded for p6spy project, and your credentials and pseudo db connection may be configured in alfresco.xml. Finally the mariaDB example for Alfresco resource has a postgresql based url parameter instead of a mysql one.
  • 9.6. Audit: In the logstash client, the information is consumed via curl command pointing to the audit webscripts, so it may be necessary to change the credetials inside the shell scripts, to make it work.


Active Member II

Hi Cesar,

This tool does not need much in term of resources but it does always depends on the customer environment i.e. how many nodes do they have sending information to Elasticsearch? From experience I have worked with customers that deploy Elasticsearch on a VM machine with 8 CPUs and 16 Gb of ram but since it is a java application you can limit how much memory needs to use using standard JVM options i.e. -Xmx

DB queries for workflows and activities should be fine since I'm querying just one can always limit how often these queries are executed by changing the configuration in the logstash.conf file.

Alfresco already provides the Support Tools (see What's new in Alfresco Content Services | Alfresco Documentation). One major difference is that the data in the support tools is not persisted, meaning one you move away from the web page the data disappears, while with Elasticsearch the data is stored for as long as you decide.

I believe P6SPY will only log data if the slow query threshold has been is important to define one otherwise all queries are written to file and Alfresco executes far too many to log them all.

I see...mysql creates the table names in capitals...I will update the blog to mention that...thank you!

You should be able to reach kibana at localhost (on the server running docker) since the port is being exported to the server. Of course if you are on a different machine to the one running docker then you will need to use the IP address of the server running docker.

Thank you for pointing out the typo in the MariaDB configuration...I will update that too.

You are right about the audit script...I will add a note to change the login there.

Thank you for all your comments....much appreciated.




Hi Miguel:

Hope you are doing well. Some minor questions, back on this topic:

- Why is Redis needed in the monitoring picture ? Can you simply send data from logstash client to elasticsearch ? 

- It seems that anything is based in JMX. Is it correct ? So it can be applied to Alfresco Community completely (last time I checked with an Alfresco EE setup).



Active Member II

Hi Cesar,

Redis is just acting as a broker to store data to be read by logstash...not a must have component but it is handy when logstash is down for maintenance, etc.

That's right, there is no JMX data so it should work with Alfresco community version.




Thanks again Miguel for your feedback, this is very useful. 



Active Member II

Hi Miguel:


I have some minor questions too, 


- Can this do some action on sone perticular log like shooting a mail to tell if there was a heap space error or any 401 response

- Can be applied to Alfresco Community completely?


Active Member II

Hi Tushar,

I didn't implement actions on events entries (alerts) from logs but that can be easily implemented. Take a look at this project for further information: GitHub - Yelp/elastalert: Easy & Flexible Alerting With ElasticSearch 

And yes, it should work with Alfresco community.



Active Member II

We have followed all the steps mentioned above up until #10 but couldn't get it working. We could see the server performance stat but not the Alfresco Log Analyzer dashboard.

Any help is highly appreciated.

Probably LogStash is not recognizing the logs generated by Alfresco.