Platform Limitations

Liferay Cloud and its services have some notable limitations, depending on your subscription level. Limitations with Liferay Cloud’s infrastructure may change over time.

Overview

Take these general limitations into consideration when planning to use Liferay Cloud:

  • Limits apply on the available vCPUs, memory, scaling, network configurations (domains, SSL certificates, and IP addresses), and VPN bandwidth for each service. For instance, each service is limited to a maximum 200 GB RAM. Custom domains are also limited to 50 or 1500, depending on your web server’s configuration.

  • Concurrent operations (such as concurrent uploads), build size, concurrent builds, and backups also have limitations.

  • Service downtime may occur due to planned maintenance, most notably for environments using a single instance of the Liferay or Search services.

  • A private cluster subscription may be needed for more stringent security, compliance, or VPN requirements.

See the further sections below for more details.

All Services

These limitations apply to every service in a Liferay Cloud environment:

Access to old logs: By default, logs from the last 30 days for each service are available via the console. Submit a Support request to access older application logs (up to a year old). Status and build logs cannot be retrieved after 30 days.

Additional instances per service: The scale setting initially uses the purchased number of instances for your subscription plan. By default, services can add one additional instance, with the exception of Search. The Search service must use an odd number of instances.

Downtime: Services running with a single instance may experience restarts when the Liferay Cloud infrastructure is updated for scheduled maintenance. Use high availability settings (two instances each of the Web server and Liferay services, and three instances of the Search service) for production-type environments to avoid disruptions. You can view the schedule for planned maintenance here.

Memory per service instance: Services can have up to a possible 200 GB of RAM, and this is determined by your subscription plan. The default plan has 16 GB per service.

Virtual CPUs per service instance: Services can have up to a possible 32 vCPUs, and this is determined by your subscription plan.

Liferay Service

These limitations apply to the Liferay service in each Liferay Cloud environment:

Auto-scaling: When enabled, auto-scaling may only add new instances up to a default maximum of 10. You can configure a different maximum number of instances, up to 100 instances.

Document Library Storage: Your subscription plan determines the size of the data volume for the Liferay service; this includes storage used for Liferay’s Document Library. The default volume size is 100 GB, but it can be increased as long as the size is below 4 TB. A private cluster is required if a project needs more than 4 TB of storage.

Session Replication: Replicating sessions between multiple Liferay instances in Liferay Cloud may impact your instances’ performance, and is not supported. Liferay Cloud leverages pod evictions on Kubernetes as part of the resource management and high availability strategies. Pod evictions are necessary to maintain the health of the system. They are triggered by scenarios such as updates to the underlying infrastructure or when nodes approach their resource limits. This results in the eviction of active user sessions from the affected pods. Because sessions are lost, session replication causes performance issues, and may also cause users unexpectedly to be logged out. Instead, use sticky sessions, or avoid using session storage entirely in your custom applications.

Note

Liferay mitigates the effects of session loss through various strategies such as cookies and database interactions. This ensures items like Commerce carts are saved even when the session is lost.

Dynatrace

Dynatrace is not included in the Standard setup for Liferay Cloud environments, but it can be purchased separately to use with it. Dynatrace is included in the High Availability setup, but only for Production or UAT environments.

These limitations apply to Dynatrace:

Dynatrace Metrics Discrepancy: Dynatrace metrics do not match the metrics shown in the Liferay Cloud console. This is because Dynatrace collects process-level metrics directly from the host using its agent-based monitoring, while the console metrics measure the pod-level resource quotas and limits within Kubernetes. So both may have graphs with different metrics for the same resources (e.g. CPU) and they are both correct for different contexts within the application.

Streaming Liferay Logs: Liferay logs cannot be streamed to Dynatrace logs.

Session Replay: The Dynatrace Session Replay feature is not available with Liferay Cloud.

Database Service

These limitations apply to the Database service in each Liferay Cloud environment:

Database Size: The size limit for a database container is 4 CPU cores, 15 GB of memory and 100 GB of disk storage. You can increase these limits with add-ons to your subscription.

Downtime: Database maintenance may cause downtime every few months. This downtime usually lasts about two minutes. This may not come with a notification in advance. You can set a preferred maintenance window to mitigate the impact of downtime.

Read/write splits: Configuring a read/write split in your database service is not supported in Liferay Cloud.

Search Service

These limitations apply to the Search service in each Liferay Cloud environment:

Configuration: Elasticsearch must be configured through the Liferay Cloud workspace, and not the Liferay UI. The configuration file in the project workspace is used on each deployment and overwrites the previous deployment’s configuration.

Memory Settings: The default (and maximum) JVM heap size for the Elasticsearch server is 4 GB. The maximum allocation is determined by your subscription plan.

OS Packages: Installing additional OS packages for the Search service is not supported.

Pod Management Policy: Elasticsearch nodes in a cluster must connect to each other to start successfully. For search services with multiple instances, the podManagementPolicy value in the service’s LCP.json file must be set to parallel to avoid issues with the service starting up.

Backup Service

These limitations apply to the Backup service in each Liferay Cloud environment:

Backup Consistency: As with any process copying from a database with changing data, consistency between data in the database and document library cannot be guaranteed if a backup is created while updates are occurring. To ensure a completely consistent backup, coordinate with your database administrator to freeze updates while you perform a manual backup.

Backup Size: Before Liferay Cloud version 4.2.0, backups used ephemeral storage. The size of backups in these versions is limited to the remaining space on a shared ephemeral disk, which may vary.

Backup Uploads: Only one backup may be uploaded per minute.

Concurrent Operations: Concurrent backup creation, restores, or uploads are not supported. However, concurrent downloads are supported.

Resource Allocation: The RAM and number of vCPUs allocated to the Backup service are determined by your subscription plan. The default allocation is 2 vCPUs and 1 GB of RAM for the service.

Upload/Download Speed: The speed of backup uploads or downloads is limited by your internet connection speed and the size of the backup. It may take up to several hours to download a backup with a very slow connection.

Web Server Service

These limitations apply to the Web server service in each Liferay Cloud environment:

Plugins: Installing additional plugins for the web server is not supported.

Resource Allocation: The web server has 2 vCPUs and 512 MB of memory by default. This may result in slower response times for large uploads or downloads. Your subscription plan determines the specific resource allocation for the service.

Continuous Integration Service

These limitations apply to the CI service in each Liferay Cloud environment:

Administrative access: Admin-level access is not allowed on the Jenkins server. Instead, use the Jenkins pipeline hooks to extend the CI pipeline. Existing DevOps processes may need to be adjusted to conform to this pipeline.

Concurrent API Calls: Projects cannot perform concurrent calls to Liferay Cloud APIs. This includes tasks such as deploying a build to an environment through the CLI tool.

Resource Allocation: The RAM and number of vCPUs allocated to the CI service are determined by your subscription plan. The default allocation is 4 vCPUs and 8 GB of RAM for the service.

Server capacity: Your subscription plan determines the size of the data volume for the CI server. The default size is 100 GB.

Builds

These limitations apply to any builds created within a project:

Build Size: Individual builds are limited to 2 GB each. You must ensure that the total size of the project in your repository is less than this limit.

Concurrent Builds: A maximum of two concurrent builds may run on Jenkins because two executor threads are used.

Maximum Builds per Day: Builds are limited to 300 per day. Submit a Support request to increase this limit.

Private GitHub Servers: Integration with private GitHub servers is not supported.

Custom Services

These limitations apply to any custom services in a Liferay Cloud environment:

Host OS Access: Privileged access to the host Operating System kernel is limited to subscriptions that include a private cluster.

Security

These limitations apply to the security features available within Liferay Cloud:

Antivirus: The default Liferay DXP feature for scanning viruses on file upload cannot be used. Liferay Cloud’s Antivirus solution is used instead. Uploaded content is scanned on a schedule, and thus risks may not be detected immediately when a file is uploaded.

Authentications per Minute: A maximum of 8400 authentications are allowed per minute.

Firewall Rules: You must purchase a subscription with a private cluster and coordinate with Liferay Cloud Support to set custom firewall rules. Custom firewall rules cannot be used with a shared cluster subscription. Any custom firewall rules created for a private cluster apply to all environments in the project.

IP Address Filtering: IP address filtering can only be applied on the web server service.

File Storage

These limitations apply to file storage for multiple services:

Ephemeral Storage: Ephemeral Storage is used for all files not stored in volumes. Ephemeral Storage is located on the host node’s internal storage, and it is shared between all containers running on that node. If a container requests more space than the host node has available, then the container is moved to another node. The hosts disks have a capacity of 250 GB.

Sharing Data Between Services: Services with the StatefulSet Deployment Type cannot share data with other services.

StatefulSet Storage Size: You must make a Support ticket to add storage for services with the StatefulSet Deployment Type. The storage size of StatefulSet services cannot be reduced once it is increased.

Network Configuration

These limitations apply to the network configuration of your services in a Liferay Cloud environment:

Changes to Custom Domains: There may be a delay (up to 60 minutes) for changes or additions to custom domains to propagate.

Maximum Custom Domains: There is a limit of 50 custom domains if you have multiple services exposed outside of the environment (in addition to the default web server). However, the web server can use a limit of 1500 custom domains if it is the only point of entry. You must open a Support ticket if you change your web server service to increase the usage beyond this limit.

Maximum SSL Certificates: A maximum of 14 custom SSL certificates are allowed. The provider issuing the certificates may also impose its own limitations to make this less.

Outbound Connections: Connections from the project environment to external endpoints use a NAT solution with a 120-second timeout per connection. New connections created successively in a short period of time can lead to port exhaustion and dropped outbound traffic. Connection pooling can prevent this problem.

Public IP Addresses: By default, every environment has one public IP address, and services within the environment have internal IP addresses. However, you can configure a service’s ports to be external, assigning a public IP address to the service. Exposing a service’s endpoints to the internet is not recommended, because it bypasses DDoS protection from the HTTPS load balancer.

Wildcard SSL Certificates: Wildcard certificates are not supported for Liferay’s auto-generated SSL certificates. However, you may configure your instance with custom Wildcard SSL certificates.

TCP Port Status: Some TCP ports may show as open, even if no external traffic can reach the customer environment using it. These ports are open for other projects sharing the same disk on the server, and do not pose a security risk for your environment.

JMX Exporter Port: Port 15000 is used by the JMX Exporter and can not be used for any other purpose.

VPN

These limitations apply if you have connected a VPN server to the services in your environment:

Site-to-Site VPN: Site-to-Site VPN servers can only be configured with Google Cloud VPN. It also requires a private cluster subscription.

Bandwidth: Each VPN tunnel is limited to a total bandwidth of 3 Gbps. This limit applies to the total of incoming and outgoing traffic.

Extending On-Premises Networks: Cloud Interconnect or Express Route dedicated connections from an on-premises network are not supported. This applies to connections made directly or via partner providers, and with shared or private clusters.

Two-Factor Authentication: Using two-factor authentication within your Cloud VPN (OpenVPN or IPSec) is not supported.

Capabilities

Product

Contact Us

Connect

Powered by Liferay
© 2024 Liferay Inc. All Rights Reserved • Privacy Policy