Temporal Cloud's new `temporal_cloud_v1_billable_action_count` metric exposes per-second billable action rates broken down by Workflow Type and Action Type. The post walks through practical PromQL queries for cost analysis (identifying top Action consumers, estimating 24-hour totals), debugging (detecting rare code paths and deployment regressions by observing unexpected Action Type spikes), and alerting (ratio-based spike detection, retry-rate thresholds). Operational notes cover query conventions (avoid rate()/increase()), 3-minute data latency, API rate limits (180 req/hour), and cardinality management for the high-cardinality metric. The metric is explicitly not a billing substitute but is the most granular near-real-time tool for understanding Action consumption patterns.

10m read timeFrom temporal.io
Post cover image
Table of contents
What this metric is, and isn’t #Usage optimization analysis #Debugging Workflow execution paths #Alerting #Other ways to use this metric #Operational reminders #Summary #Related resources #
52 Impressions