Updated metrics pricing for Elastic Observability: Best-in-class metrics — now cheaper, too!

Updated metrics pricing for Elastic Observability: Best-in-class metrics — now cheaper, too! We’re slashing the price of TSDS metrics on Elastic Observability Serverless by 75% By Vinay Chandrasekhar June 29, 2026 Share on Twitter Share on Twitter Share on LinkedIn Share on LinkedIn Share on Facebook Share on Facebook Share by Email Share by Email Print this page Print As part of the significant product enhancements being announced around metrics, Elastic Observability is updating how it prices metrics. On Elastic Observability Serverless, metrics data stored in TSDS index mode is now priced at 25% of the standard Observability per-GB rate for both ingest and retention. On Elastic Cloud Hosted (ECH) and Self-Managed, metrics on TSDS is available for no additional charge at this time. On both, metrics now simply require less infrastructure to run. These pricing changes accompany the recent revamp of how Elastic Observability stores and queries metrics. Elastic rebuilt the metrics platform around a columnar storage engine purpose-built for time series. It stores metrics up to 2.5x more efficiently and reduces query latency by up to 30x versus competition. It also adds native Prometheus ingest and PromQL support, along with comprehensive ES|QL time series querying support. Those engineering improvements make Elastic incredibly efficient to run for metrics, which means customers now pay a fraction of the costs compared to other vendors. This blog covers what is changing and then walks through an illustrative cost comparison with Datadog where Elastic delivers dramatic cost savings, well over 50%. Elastic Observability Serverless: Metrics now priced at 25% of the standard per-GB rate On Elastic Observability Serverless, TSDS metrics data is now priced at 25% of the standard Observability per-GB rates for both ingest and retention. How you are billed does not change. You still pay per GB ingested and per GB retained on the same volume-tiered structure you already use. TSDS index mode metrics are simply priced at one-quarter of the rate that applies to other observability data at each tier. In practice, that means metrics in Observability Complete are priced approximately as follows: Dimension Standard Observability rate (as low as) Metrics rate (as low as) Ingest (per GB) ~$0.09 ~$0.023 Retention (per GB/month) ~$0.019 ~$0.005 Rates shown are the deepest volume-tier ("as low as") prices. Entry-tier rates are higher and step down with volume. Metrics rates are 25% of the corresponding standard rate at every tier. Pricing applies to metrics stored as time series (TSDS) on Elastic Observability Serverless. There are a few things are worth calling out: You keep paying for what you use in units you already understand. There is no separate metrics billing model to learn, no per-host license, and no cardinality-based surcharge. A new Kubernetes label, a new cloud instance tag, or a new application dimension adds data, but it does not move you into a different or custom pricing model. Long retention stays inexpensive. Because metrics retention is priced at 25% of the standard rate, keeping a year or more of metrics costs a fraction of what signal-agnostic pricing would charge. Shorter retention windows cost proportionally less. Volume pricing tiers still apply. As your ingest and retained volume grow, your per-GB unit price decreases across the same tier structure used for the rest of observability. Elastic Observability Serverless price comparison with Datadog Datadog prices infrastructure monitoring primarily per host and then bills separately for custom metrics: any user-defined metric beyond Datadog's built-in integrations, which might include OpenTelemetry, Prometheus, or Kubernetes data. As instrumentation gets richer and cardinality grows, the custom-metric and container line items climb on top of the per-host charge. Elastic prices the same workload on data volume at the metrics rate shown above. To make the comparison concrete, here are two illustrative workload examples. Both use conservative, clearly stated assumptions. Shared assumptions Datadog:  Infrastructure Pro at its lowest published (annual commit) host price of $15/host/month, which includes 15-month retention Custom-metric overage at the $5 per 100 metrics/month list rate Container overage at $1/container/month Elastic Observability Serverless:  Observability Complete Premium support 15-month retention to match Datadog's included 15 months Note that storing data for less than 15 months will lead to lower prices with Elastic. Both:  15-second collection interval (Datadog's Agent default applied to both platforms) ~450 bytes per datapoint (approximate based on test data at ingestion across OpenTelemetry and Prometheus on Kubernetes and hosts) Figures are rounded monthly estimates for illustration Example 1: A 300-host cloud and VM fleet Consider a team monitoring 300 hosts — each emitting roughly 500 metric series, including about 250 custom application metrics per host (150 above Datadog's included allotment of 100 per host).   Datadog (Infrastructure Pro*) Elastic Observability Serverless Hosts $4,500 (300 × $15) — Custom-metric overage $2,250 (250 − 100) × 300 hosts = 45,000 over allotment; 45,000 ÷ 100 × $5 — Metrics data (ingest + 15-month retention) Included ~$2,200 Ingest $693.22 + retention $1,240.10 = $1,933.32 subtotal. Premium support ×15% = $290.00 Estimated monthly total ~$6,750 ~$2,200 Elastic comes in about 67% below the Datadog estimate for this illustrative workload. Example 2: A Kubernetes platform (5 clusters, ~200 nodes) Consider a higher-cardinality Kubernetes environment: 200 nodes, ~1,000 series per node, ~400 custom metrics per node, and about 20 containers per node.   Datadog (Infrastructure Pro*) Elastic Observability Serverless Hosts $3,000 (200 × $15) — Custom-metric overage $3,000 (400 − 100) × 200 nodes = 60,000 over allotment; 60,000 ÷ 100 × $5 — Container overage $3,000 (20 − 5) × 200 nodes = 3,000 over allotment; 3,000 × $1 — Metrics data (ingest + 15-month retention) Included ~$2,800 Ingest $835.21 + retention $1,618.47 = $2,453.67 subtotal. Premium support ×15% = $368.05 Estimated monthly total ~$9,000 ~$2,800 Again, Elastic lands about 69% below the Datadog estimate. *A note on the Datadog tiers: Infrastructure Pro at $15/host/month, the lower-priced and more common tier, is shown above. Datadog Enterprise costs more per host ($23) but includes larger allotments of 200 custom metrics and 10 containers per host, versus 100 and 5 on Pro, so the totals tend to land in the same range. In Example 1, Enterprise works out to about $7,650 per month (Elastic roughly 71% lower) versus Pro's $6,750 (roughly 67% lower); in Example 2, Enterprise works out to about $8,600 (Elastic roughly 67% lower) versus Pro's $9,000 (roughly 69% lower). The comparison reaches the same conclusion on either tier. What the comparison shows: Cost savings of well over 50% In both examples, Elastic Observability Serverless costs roughly two-thirds less than the comparable Datadog bill with retention matched at 15 months on both sides. The gap is driven by how each platform prices metrics: Datadog's custom-metric and container charges accumulate on top of the per-host fee as instrumentation grows while Elastic bills on data volume alone. The size of any individual saving depends on cardinality, collection interval, and retention. These examples use Datadog's 15-second default collection interval. At a 60-second interval or with lighter custom-metric and container usage, the percentages shift, but the direction does not. The conclusion holds across a wide range of assumptions: For metrics, Elastic Observability Serverless is well over 50% cheaper than Datadog, and you keep every metric at full resolution rather than dropping data to manage cost. Frequently asked questions How are metrics billed on Elastic Cloud Serverless? Metrics are billed the same way as the rest of Elastic Observability: per GB ingested and per GB retained on a volume-tiered structure. With this update, metrics data is priced at 25% of the standard Observability per-GB rate on both dimensions, which works out to roughly $0.023 per GB ingest and $0.005 per GB per month retention at the highest volume tiers. Does Elastic charge more for high-cardinality metrics? No. Elastic does not use cardinality-based or per-metric billing. Adding a new Kubernetes label, cloud instance tag, or application dimension adds data volume, but it does not trigger a premium tier or custom-metric surcharge, so you can keep high-cardinality metrics at full resolution. Is Elastic Observability cheaper than Datadog for metrics? In the illustrative workloads in this post, Elastic Observability Serverless costs about two-thirds less than Datadog for metrics, modeled at Datadog's 15-second default collection interval with retention matched at 15 months on both sides. Datadog bills primarily per host plus custom-metric and container overage while Elastic bills on data volume, so the gap widens as cardinality grows or collection frequency drops. Actual savings depend on your hosts, cardinality, collection interval, and retention. How much does it cost to keep a year or more of metrics on Elastic? Because retention is priced at 25% of the standard Observability rate (as low as ~$0.005 per GB per month), long-window metrics retention on Elastic Cloud Serverless is still extremely reasonably priced, and allows you the control to choose and pay for your retention period. In the comparisons above, 15-month retention is shown in the Elastic figures. Shorter retention windows cost proportionally less since you pay per GB retained. Why are Datadog metrics bills high for Kubernetes and OpenTelemetry workloads? Datadog classifies most OpenTelemetry, Prometheus, and Kubernetes app metrics as custom metrics, which are billed on top of the per-host charge once they exceed the included allotment, alongside container overage. High-cardinality, densely instrumented environments cross those thresholds quickly. Elastic prices the same data on volume alone, which is why the cost difference is largest for exactly these workloads. Availability and getting started The new metrics pricing accompanies the broader Elastic Observability metrics release, which includes the columnar metrics engine, native Prometheus ingest, PromQL support, ES|QL time series queries, prebuilt Kubernetes and AWS content, and migration tooling for Datadog and Grafana. Start a free trial Create a pricing estimate Elastic Observability documentation Pricing figures, competitive comparisons, and example workloads in this post are illustrative estimates based on published list pricing as of the date of publication. Actual costs vary by workload, configuration, retention, and commercial terms. Competitor pricing is based on publicly available information and is subject to change by the respective vendor. The release and timing of any features or functionality described in this post remain at Elastic's sole discretion and may change. SHARE Share on Twitter Share on Twitter Share on LinkedIn Share on LinkedIn Share on Facebook Share on Facebook Share by Email Share by Email Print this page Print