Welcome to the Hazelcast Management Center Release Notes. This document includes the new features, enhancements and fixed issues for Hazelcast Management Center releases.

Please select a release from the Table of Contents on the left panel to see its release notes.

Hazelcast Management Center enables you to monitor and manage your nodes that are running Hazelcast. In addition to monitoring the overall state of your clusters, you can also analyze and browse your data structures in detail, update map configurations, and take thread dump from nodes. With its scripting and console module, you can run scripts (JavaScript, Groovy, etc.) and commands on your members.

During development, Management Center provides deep insights. In production, it can be directly used by IT operations or integrated with enterprise monitoring tools using technologies, such as REST and JMX. Management Center also provides dedicated controls for WAN Replication for monitoring replication, and to manipulate services, configuration, and processes.

Management Center provides the fundamental properties of the selected cluster with an easy to use interface.

The numbers in the square brackets refer to the issue number in Hazelcast Management Center GitHub repository.



  • Detail screens for data structures was not working on Internet Explorer and Edge. This is fixed by adding the related compatibilities. [#1301]

  • Fixed an issue where some of the graphs were showing multiple lines. [#1273]

  • Fixed an issue where the monitoring screen for maps were becoming unresponsive after some time. [#1252]

  • Fixed an issue where the graph states were not refreshing when switching between the clusters, i.e., state from the previous cluster was shown. It was also occurring when switching between the instances of same data structures within one cluster. [#1250]

  • Fixed an issue where the "Rolling Upgrade" and "Hot Restart" pages of the "Administration" menu was not shown according to the licenses in hand. [#1248]

  • Fixed the overlapping of memory and CPU utilization parts of the "Status" screen. [#1232]

  • Fixed various cosmetic issues related to distributed object graphs and statistics. [#1217]

  • Fixed the incorrect representation of the "Value" column in a specific PN Counter page. It was showing the difference between the number of add operations and subtract operations instead of the current value of the counter on the related member. [#1203]

  • Fixed the incorrect screen message when Management Center is started with Hazelcast Enterprise license and no cluster is selected, e.g., after login. In this case, the "Rolling Upgrade" and "Hot Restart" pages of the "Administration" menu was showing the message "You don’t have a license to access this URL". [#1195]

  • Fixed an issue where the graphs for map and cache structures was showing data from previous statistics. [#1145]


New Features

  • Introduced "License" screen which shows the details of your Hazelcast Management Center and cluster licenses. See the License Information chapter for more information.

  • Introduced Delta WAN Synchronization which uses Merkle trees. See WAN Sync section for more information.


  • Content of the thread dumps on the Members screen can now be copied to clipboard. [#1125]

  • Hot Restart screen now shows the directory which is configured for Hot Restart backup. [#1035]

  • Upgraded to the latest Jetty version to eliminate the security vulnerabilities. [#989]

  • Improved WAN synchronization screen so that now it shows a stream of events that happen on each member. [#477]

  • Improved WAN replication monitoring screen by making the WAN replication status and operations for each member more obvious. [#316]

  • WAN replication could only be stopped on a per-member basis. The option to stop it on the wan-replication level has been added so that all members can stop replicating. [#314]


  • Fixed the Management Center user interface for Internet Explorer 11 on Windows 10. It was showing only the logout button. [#1207]

  • Fixed an issue where long string values of a map were not wrapped in the Map Browser screen. [#1172]

  • Fixed incorrect units shown in some of the statistics graphs. [#1146]

  • Fixed uptime information shown for clients which was "0" even the clients were running. [#1120]

  • Fixed WAN synchronization error message which read as "undefined". [#1105]

  • Fixed an issue where expanding the WAR file was causing Management Center to crash. [#1028]

  • Fixed an issue where changing to Hot Restart tab was not updating the browser tab title. [#1005]

  • Fixed an issue where the names for some maps were not visible on the Map Memory Distribution screen. [#853]

  • Hot Restart Backup operation was not showing any progress. This is fixed by migrating the frontend component to React. [#758]

Behavioral Changes

  • Since the system property hazelcast.mc.max.visible.instance.count is removed with the release of Hazelcast IMDG 3.11, you may see more than usual data in your Management Center, if you configured the system property and have more than the specified amount of data structures in your Hazelcast cluster. You can disable statistics for individual data structures in your Hazelcast IMDG configuration to control the amount of data in your Management Center.

  • When TLS is enabled in standalone mode (via command line argument -Dhazelcast.mc.tls.enabled=true), HTTP port will be disabled by default. If you want to preserve the old behavior and keep HTTP port open and redirect to HTTPS port, use -Dhazelcast.mc.tls.enableHttpPort=true command line argument.



  • Added the support for configuring Management Center only with secure (TLS) port. Before, it could be configured both with non-secure (plain HTTP) and secure (TLS) ports at the same time when a secure Management Center installation is needed. See the Disabling HTTP Port section.

  • Introduced new statistics to Clustered REST and Clustered JMX APIs including average latency metrics for get, put and remove operations on the Map, MultiMap and Replicated Map data structures. See Clustered REST and Clustered JMX sections.

  • Changed the "Hot Restart Backup" text to "Hot Backup" to eliminate expressional confusions. Also changed the corresponding button icon to have it more related to its function.

  • Instead of launching the browser once you start Management Center, the starting process now prints only the URL of it so that you can paste it into your preferred browser.


  • Fixed the incorrect representations of latency values, in grid views, which are multiplied by 1000. Also "Outbound Queue Size" value representations for WAN replication has been fixed; they were incorrectly reported in KB. [#1087]

  • Fixed the slow start of Management Center on Microsoft Windows operating systems. It was caused by the extraction of WAR files. [#1072]

  • If an Enterprise license is not configured, Management Center does not show the WAN Replication page now. [#999]

  • If there is a "/" character in the name of a map, Management Center was not able to show the map monitoring page. This is fixed. [#987]

  • Fixed the 403 HTTP error when TLS is enabled. It was related to the version of Jetty and HTTP was not redirected to HTTPS. [#980]

  • Fixed the calculation of file size when the disk space is large. It was calculated to a negative value. [#964]

  • Fixed the "SNAPSHOT" text for Jetty version while Management Center is starting up. [#376]


New Features


  • "Start WAN Sync" button gets disabled unnecessarily. [#983]

  • Having multiple members on the "Status" page, the members' IP addresses should be clickable to go to the details pages of the members. [#974]

  • When the time travel is on and there are multiple clusters, it uses a single file to store all data. The content of this file prints the info related to both clusters. [#958]

  • Rolling Upgrade screen does not show the current cluster version. [#951]

  • Custom variable replacers don’t work. [#993]



  • Introduced variable replacer support that allow the users to replace values on both system and other property files that Hazelcast Management Center accepts. Please see Variable Replacers.

  • Added ability to configure different SSL keystores and truststores for different parts of the system such as the HTTP client, LDAP and Clustered JMX. Please see Enabling TLS/SSL for LDAP


Starting with 3.10, you need Java Runtime Environment 1.8 or later to run Hazelcast Management Center properly.

New Features

  • Monitoring Flake ID Generators: Introduced monitoring the number of ID batches generated by the Flake ID Generators in your Hazelcast cluster. Please see Monitoring Flake ID Generators.

  • Monitoring PN Counters: Introduced monitoring the PN Counters in your Hazelcast cluster, including the operations per second performed on and the current values of them. Please see Monitoring PN Counters.


  • Overall user interface experience has been renewed including navigation improvements and ability to bookmark pages.

  • The script startManCenter.sh can be renamed to start.sh since Management Center is a separate distribution now. [#878]

  • Logged in user should be displayed on the Management Center screen. [#820]

  • File store should be bound to a maximum usage of disk space. [#766]

  • Introduced grid views to provide an overview of all the data structures supported in Hazelcast Management Center. Before, the data structure instances were listed under each data structure type on the left menu; and this was making the navigation difficult to follow when the list is long. With this grid view enhancement, only the data structure types are listed in the left menu; when you click on a type, a new tab opens on the right listing all its instances along with related parameters. [#643]

  • On topics/reliable topics pages, when we aggregate stats for all members, we can sum up the publishes/receives in the aggregate row. [#655]

  • Non-SSL Management Center connects to an SSL cluster without problems (or will accept data from the cluster), but there should be a warning or error saying that the operations will not work. [#638]

  • Locks and dirty entries need to be removed from Replicated Map details page as they are always sent with the value 0 from the cluster members. [#634]

  • Password hints disappear too fast). [#540]

  • Introduced support for reverse proxy SSL offload. [#387]

  • Switched to the latest Jetty version (Jetty 9). [#386]

  • Management Center should highlight the "TOTAL row in data tables. [#138]


  • In any page other than the homepage, refreshing a page causes the timeline image in time travel dialog disappear. [#861]

  • Time travel slider has no effect on the data shown. [#860]

  • Hot Restart related functionalities should not be available if it is not configured. [#855]

  • Only "Logout" link is visible on Internet Explorer 11 with Windows 10. [#851]

  • "Member’s Hot Restart Status" does not show any progress. [#848]

  • Errors are not fired to the UI, the users need to check the console output or logs. [#840]

  • Buttons on the member details screen should be visible to only admin and the users with write permission. [#836]

  • When you click on the rightmost graph dropdown, items cannot be read; if you decrease the screen size, the items disappear. [#824]

  • Invalid licenses should not be allowed. [#817]

  • After logging into Management Center, it should not show a blank screen if there are not any clusters available. [#811]

  • "Running" column in the "WAN" monitoring pages is empty. [#732]

  • "Version" column in the "Clients" monitoring pages shows N/A. [#730]

  • The initial Management Center dialog for selecting security is not properly aligned in case the browser window is resized; the horizontal scrollbar is also missing. [#561]

  • When Memory Distribution graph is calculated, Native memory usage is not taken into account. If there are maps with a large native memory usage, the graph shows 100% for maps, even though there are free heap space and other space being used. [#559]

  • When a Hazelcast instance starts to use a long hostname as the public address, the partition distribution image drawing fails when the window is resized. [#482]

  • Memory distribution graph only takes the memory used by maps into account while calculating non-Hazelcast entity memory usage. It should consider all the memory usage by Hazelcast for this. [#405]

Behavioral Changes

  • Starting with 3.10, Hazelcast Management Center has a default limit for the used disk space; it will purge metrics and alert data if this limit is reached. If you want to keep these data, you have to increase the default limit and might want to change the disk usage mode to block, in which case Hazelcast Management Center will not purge any data. Note that this configurable behavior has no effect on configuration data or user accounts (they are not purged no matter the configuration is). Please see the Configuring Disk Usage Control section.

  • Starting with 3.10:

    • Hazelcast Management Center’s default URL has been changed from localhost:8080/mancenter to localhost:8080/hazelcast-mancenter.

    • Default home directory location has been changed from <user-home>/mancenter-<version> to <user-home>/hazelcast-mancenter-<version>.

    • Name of the WAR file is named has been changed from mancenter-{version}.war to hazelcast-mancenter-{version}.war.



  • The button "Clear Queue" in WAN does not work: it does nothing after the publisher is stopped. [#779]

  • Added a system property to disable TLS hostname verification while connecting to TLS-enabled clusters. [#770]

  • WAN synchronization configuration fails: after running a Hazelcast member and adding a transient WAN configuration, there is no error, but the member logs it that REST API is not enabled. [#769]

  • Except the "Size" graph in the Replicated Map’s screen, there is no statistics shown. [#714]



  • The administrator users are allowed to change other users' passwords (and also their own passwords). Now, Management Center requires the administrator users' password when they want to change other users' passwords or delete them. [#721]

  • Introduced warnings on the Management Center user interface and also loggings when thread dump and execute script requests fail. [#717]

  • To prevent brute-force password guessing attacks, repeated failed login attempts cause the login for a specific username to be disabled for some time. This behavior is now configurable. Please see Disable Login Configuration. [#697]

  • Introduced the support for configuring licenses from the command line: Management Center now reads licenses also from a system property (hazelcast.mc.license). Please see Starting with a License. [#618], [#541]


  • Multiple login issue: session expires when the same user logs in multiple times. [#668]

  • Read-only and Metrics-only users are logged out when they open the map tab. [#666]



  • Introduced a command line parameter (hazelcast.mc.forceLogoutOnMultipleLogin) to enable/disable forced logout on multiple login attempts with the same user name. [#629]

  • Password complexity rules have been hardened. [#621]

  • Introduced a global error page for unhandled exceptions. [#620]

  • Time travel is no longer allowed for the clusters whose names include non-alpha-numeric characters other than dash (-) and underscore (_). A warning is shown to the user if time travel is enabled such clusters. This enhancement was needed since by having special characters in the cluster name, one can traverse the filesystem and cause harm. [#608]

  • Management Center is now compatible with Java 9; it was not working under JDK9. [#538], [#600]

  • The session timeout has been made configurable for the embedded mode. It was by default 30 minutes. [#461]


  • Upon the very first startup of Management Center, after entering the license, when a cluster with WAN replication enabled is viewed, the "WAN" tab is not shown. It is only shown after the user logs in again. [#637]

  • Cache browser does not work for records on remote member: when Management Center tries to get a cache entry, execution fails if the record is not available locally on the member. [#603]

  • Even if the license key is provided during the initial configuration, the warning line "Node Limit Exceeded" is still shown in Management Center. [#564]

  • Management Center’s "Map Config" dialog shows "default" as the map name although it is not "default". [#501]

  • The session ID must be updated after logging into Management Center. [#463]

  • Multiple simultaneous logons must not be allowed using the same user ID as a security requirement. [#462]

  • The error messages shown to the end users are not sanitized: In these messages, Management Center should reveal only the information that the end users need to know. [#460]

  • The "Operations" dropdown menu does not show the options after a license update. [#396]

Behavioral Changes

  • Default value of the system property hazelcast.mc.max.visible.instance.count changed from 100 to Integer.MAX_VALUE. So you may see more than usual data in your Management Center, if you have more than 100 data structures in your Hazelcast cluster.



  • Introduced the display of last login time. [#575]

  • Type of the key is validated on Browse Map/Cache dialog. [#588]

  • Added extra option to start the script for the location to expand the WAR file. [#581]

  • Repeated unsuccessful login attempts cause the login to be disabled incrementally. [#578]

  • "Secure" flag is set for cookies when HTTPS is used. [#574]


  • Fix a bug where no labels were shown on "Partition Distribution" pie chart. [#570]

  • Show the correct license type when a Management Center only license is used. [#548]

  • Fix role checking for manual calls to HTTP endpoints. [#579]

  • Fix NPE during cluster restart. [#547]

  • Fix confusing WAN start/stop button. [#496]

  • Fix script’s help command. [#577]

  • Fix "No Socket Interceptor" warning for 3.8.x clusters. [#586]


New Features

  • Monitoring Clients: You can now monitor all the clients that are connected to your Hazelcast cluster. Currently only the Java clients can be monitored.


  • Notify user on incompatible member version. [#465]

  • TLS mutual authentication. [#457]

  • Promote lite members to data members. [#456]

  • Add cache browser. [#373]

  • Authentication SPI & JAAS Authentication. [#363]

  • Socket Interceptor Support: Now it is possible to manage your cluster from the Management Center if you have enabled socket interceptor on it. Please refer to Administering Management Center for more information.

  • Updated the user interface refresh period as every 5 seconds (it was every second). [#531]

  • Input on the "Change Cluster Version" dialog is now validated. [#535]

  • Introduced warning when value not found in Cache Browser. [#525]

  • Script, user and alert names now accept only alpha-numeric characters. [#511], [#514], [#515]


  • Fix security providers listed twice in the security provider dialog. [#382]

  • Incorrect High-Density Memory Store values at the Map monitoring screen. [#374]

  • Fixed cross site scripting vulnerabilities.



  • Script, user and alert names now accept only alpha-numeric characters. [#511], [#514], [#515]

  • Introduced warning when value not found in Cache Browser. [#525]


  • Fixed cross site scripting vulnerabilities.


New Features

  • Monitoring of Reliable Topic: Now it is possible to monitor your Reliable Topic data structures through Hazelcast Management Center.


  • JMX host name is now configurable.


  • WAN sync looks for incorrect map when there is more than one map in the dropdown list. [#494],[#495]

  • Inconsistent list of data structures: When there are more than two Hazelcast members, Management Center cannot show the full data structure lists every time since the members' states are updated but instance objects are not. [#488]


New Features

  • TLS Protocol Exclusions: Added an option to exclude specific TLS protocols.


  • Eliminated a security bug on the initial authentication configuration. [#459]

  • Prevented multiple notifications when changing cluster states. [#452]

  • Moved the cluster state controls to its own tab. [#451]


  • Version display is fixed so that it shows the full Management Center version including the patch release number. [#454]

  • The label of the button "Shutdown Node" in Management Center should be changed to "Shutdown Members" due to naming conventions. [#450]