azuredata-exploreranalyticscost-optimization

Azure Data Explorer cost: cluster compute plus storage

Azure Data Explorer (Kusto) bills the compute of its cluster VMs (with a markup) plus storage, running continuously unless stopped, or serverless-style with autoscale. Idle clusters and over-provisioning drive the bill. Here is the model.

The C3X Team··4 min read

Quick answer

Azure Data Explorer (ADX, Kusto) bills the compute of its cluster VMs plus an ADX markup, plus storage for the data, running continuously while the cluster is up. Cost scales with cluster size and uptime, so right-sizing the cluster, enabling autoscale to match query load, and stopping or scaling down idle clusters (especially non-production) are the main levers.

Azure Data Explorer is a fast analytics service for large volumes of telemetry, logs, and time-series data, using the Kusto query engine. It runs on a cluster of VMs, and its cost is that compute (with an ADX markup) plus storage, billed while the cluster runs, so cluster sizing and uptime are the cost decisions.

Cluster compute plus storage

ComponentBills for
Cluster computeVM cost + ADX markup, while running
StorageData stored (hot cache + persistent)
MarkupADX service fee on top of VM cost

You provision a cluster of VMs sized for your query and ingestion load, and pay their compute cost plus an ADX markup while the cluster runs, plus storage. A running cluster bills continuously whether or not it is queried, so an oversized or idle cluster wastes money, similar to other cluster-based analytics.

Autoscale and stop idle clusters

ADX supports autoscale, adjusting cluster size to load, which avoids running peak capacity constantly. Clusters can also be stopped when not needed (non-production, or workloads with idle periods), which stops the compute billing. Right-sizing the cluster and using autoscale and stop/start are the primary levers, since the compute dominates.

Controlling Data Explorer cost

Right-size the cluster to real query and ingestion load, enable autoscale so it matches demand rather than running peak constantly, stop or scale down clusters during idle periods (especially non-production), manage the hot cache size (data in hot cache is faster but costs more) versus cold storage, and set retention to what you query. Because the cluster compute dominates and bills while running, sizing and uptime discipline is the whole game, the same as Synapse dedicated pools.

FAQ

How is Azure Data Explorer priced?

By the compute of its cluster VMs plus an ADX markup, plus storage for the data, billed while the cluster runs. Cost scales with cluster size and uptime, so an oversized or idle cluster wastes money. It also has a hot-cache-versus-cold-storage tradeoff, with hot cache faster but more expensive.

How do I reduce Azure Data Explorer cost?

Right-size the cluster to real query and ingestion load, enable autoscale so it matches demand rather than running peak constantly, stop or scale down clusters during idle periods (especially non-production), manage the hot cache size versus cold storage, and set retention to what you query. Compute dominates, so sizing and uptime are key.

Does Azure Data Explorer charge when idle?

Yes, while the cluster is running. A cluster bills for its VM compute plus markup continuously whether or not it is queried, so an idle running cluster wastes money. Stopping clusters during idle periods (non-production, or workloads with idle windows) stops the compute billing, and autoscale reduces size under low load.

What is the hot cache in Azure Data Explorer?

The portion of data kept on the cluster's fast local storage for low-latency queries, as opposed to cold data in cheaper persistent storage. More hot cache means faster queries at higher cost. Sizing the hot cache to the data you query frequently, and letting older data stay cold, balances query performance against cost.

How do I right-size a Data Explorer cluster?

Match the cluster VM size and count to real query concurrency and ingestion throughput, using autoscale to handle variable load rather than provisioning for peak constantly. Monitor cluster utilization and adjust, and stop clusters during idle periods. The compute dominates the bill, so sizing it to actual demand is the main lever.

Does C3X estimate Azure Data Explorer cost?

C3X prices the Data Explorer cluster and its VM configuration from your Terraform, so the provisioned compute cost appears before deploy. Actual sizing depends on query and ingestion load, which you tune from monitoring, and storage scales with data volume.

What to do next

Size your analytics cluster before you deploy. C3X reads your Terraform and prices your resources against a live catalog. Start with the quickstart.

Try C3X on your own Terraform

Free and open source. No API key required. One command to install, one command to estimate.