Track Kafka storage and partition costs without Gateway using Conduktor Chargeback. Measure resource usage per cluster to allocate infrastructure costs.
Chargeback without Gateway lets you visualize the cost allocation based on two key metrics:
Bytes stored: the amount of data stored across all topic partitions (measured in GB-hours).
Partition count: the number of partitions maintained (measured in partition-hours).
Costs are calculated by multiplying these accumulated metrics by your configured rates (cost per GB per hour for storage, cost per partition per hour for partitions).You can view data from the perspective of overall cost, storage or number of partitions.
Overall cost: displays the total dollar cost by adding together the storage and number of partitions
Storage: shows the bytes stored, also showing the cost of said bytes, and what percentage this cost makes up of the total storage cost across all items for the selected period
Number of partitions: shows the partition count, also showing the cost of said partitions, and what percentage this cost makes up of the total number of partitions cost across all items for the selected period
You can view cost data at different levels of granularity:
Cluster: view costs grouped by Kafka cluster
Application: view costs grouped by application
Application Instance: view costs grouped by application instance (filtered by selected cluster or application)
Topic: view costs grouped by topic (filtered by selected cluster or application instance)
The measured period can be changed from 7D, 30D, 6M, 12M or YTD (Year To Date). The view can be changed per cluster, application, application instance or topic using the view dropdown.You can see the topic cost per hour and cost of bytes stored per GB hour for each cluster connected to Console.You can easily update the cost rates for one or more clusters in this same section.
Once metrics are aggregated for the selected time period, Console calculates costs using the configured cost rates.Cost:If cost is not set, will be defaulted to 0. The maximum allowed cost for each is 9.99999999.
Storage
Number of partitions
Overall Cost
(total_byte_hours / 1,073,741,824) × bytes_stored_per_hour_cost// Where `1,073,741,824` is 1 GB in bytes (1024³).
Example: A cluster with 7,200 partition-hours over a 31-day period (744 hours):
Average: 7,200 / 744 ≈ 9.68 → Average of ~10 partitions throughout the period
The accuracy of metrics depends on the collection frequency. A shorter interval (e.g., 30 minutes) captures more granular usage patterns, while a longer interval (e.g., two hours) may miss short-lived spikes.The default frequency is one hour.Note that deleting a cluster will remove related Chargeback metrics.
To facilitate external analysis and reporting, the data shown in the table can be exported as a CSV file.To export the data, click Export all at the top right of the graph.The default exported file name will be Chargeback_<startDate>_<endDate>_<activeTabName>.csv.
