Fair Usage Limits for Camunda 8 SaaS Starter & Professional Plans
At Camunda, we strive to ensure consistent availability and reliability to all customers. To improve our service capability and operation efficiency, we have some technical limits in place that apply only to Starter & Professional plan customers.
For information on the limits for the Enterprise plan, please contact us.
Organizational Limits
| Technical Limits | Starter plan |
| Maximum general users | 100 |
| Production cluster reservations | 2 |
| Development cluster reservations | 3 |
| Total clusters reservations | 5 |
Cluster Usage Limits
| Metrics* | Starter - Development Cluster | Starter - Production Cluster |
| Max Throughput Tasks/second | 10 | 65 |
| Max Throughput Process Instances/day | 100 k | 0.5 M |
| Max Total Number of Process Instances stored (in Elasticsearch in total) | 10 k | 100 k |
* The numbers in the table where measured using Camunda Platform 8 (version 8.0) and the benchmark project. It uses a ten task process. To calculate day-based metrics, an equal distribution over 24 hours is assumed.
Data Retention Policies in days
| Camunda Component | Starter - Development Cluster | Starter - Production Cluster |
| Operate | 7 | 30 |
| Optimize | 30 | 180 |
| Tasklist | 7 | 30 |
| Zeebe | 7 | 7 |
Availability Targets & Support Levels for Clusters
Availability targets and support level 1 apply to Production clusters only.
To find out more about our support, see the support overview page and the support guide.
Cluster Version Upgrade
Unsupported cluster versions are scheduled for automatic updates within 6 weeks after the end of the support date. Org owners and admins are notified via email and in-app of the end of the support date and have 6 weeks to upgrade their clusters ahead of the auto-upgrade date.
Cluster Auto-Pausing Policy
| Starter - Development Cluster | Starter - Production Cluster |
| If inactive for 2 hours or no process model is deployed Customers can easily unpause the cluster in the Console by clicking a button in the Console or by deploying a new process model to the paused cluster. | If no process model is deployed Customers can easily unpause the cluster by deploying a new process model to the paused cluster |
