Trace Sampling Cost Calculator

Calculator Inputs

Use realistic averages; then tune sampling and retention to fit budget.

Clear Saved Runs

Incoming traces per second

Peak-to-average matters; use a steady monthly average.

Sample rate (%)

Effective sampled traces = traces/s × sample rate.

Average spans per trace

Often 10–40 spans for web services; more for microservices.

Average span size (bytes)

Includes attributes, events, links, and context.

Compression ratio (stored/raw)

0.35 means 35% of raw size after compression.

Retention (days)

Average storage assumes steady ingest and rolling retention.

Currency symbol

Example: $, €, £, PKR.

Ingestion cost per million spans

Applies to sampled spans. Use your vendor rate.

Trace storage cost per GB-month

Charged against average stored GB for retention window.

Indexing Options

Indexing increases storage footprint, but improves search/query speed.

Include indexing

Indexed data (%)

Portion of trace data that becomes searchable index.

Index overhead factor (x)

Index size multiplier vs. indexed trace data.

Index storage cost per GB-month

Often higher than raw storage depending on backend.

Query / Egress Options

Optional estimate for data scanned/exported during queries or exports.

Include egress

Monthly egress (GB)

Add exports, dashboards, and cross-region traffic estimates.

Egress cost per GB

Use your cloud/network pricing rate.

Fixed monthly platform fee

Optional base subscription or support plan fee.

Tip: reduce costs by lowering sample rate, span size, retention, or indexed percent.

Formula

Example Data Table

Sample scenarios to illustrate how sampling and retention impact monthly cost totals.

Scenario	Traces/s	Sample %	Retention (days)	Index %	Est. Notes
Conservative	120	10%	7	25%	Lower cost; suitable for baseline service health.
Balanced	250	20%	14	40%	Good coverage for common incident triage.
Deep-dive	600	50%	30	70%	Higher cost; best for intensive performance work.

Run each scenario above in the calculator to produce a cost breakdown.

Formula Used

This calculator estimates steady-state monthly costs using a constant average ingest rate.

Sampled traces/s	incoming_tps × (sample_rate ÷ 100)
Sampled traces/month	sampled_traces/s × seconds_per_month
Sampled spans/month	sampled_traces/month × spans_per_trace
Raw GB/month	(sampled_spans/month × span_size_bytes) ÷ 1024³
Stored GB/month	raw_gb/month × compression_ratio
Avg stored GB	(stored_gb/month ÷ days_per_month) × retention_days
Ingestion cost	(sampled_spans/month ÷ 1,000,000) × ingest_cost_per_million_spans
Trace storage cost	avg_trace_storage_gb × trace_storage_cost_per_gb_month
Index storage (optional)	avg_index_storage_gb × index_storage_cost_per_gb_month
Egress cost (optional)	egress_gb_month × egress_cost_per_gb
Total monthly cost	ingest + trace_storage + index_storage + egress + fixed_fee

Note: seconds_per_month uses an average month length (30.4375 days) to smooth estimates.

How to Use This Calculator

Enter your average incoming traces per second and a target sampling rate.
Set spans per trace and span size using telemetry or vendor estimates.
Choose a compression ratio and a retention period that match policy.
Fill in ingestion, storage, index, and egress pricing from your provider.
Click Calculate to see totals and a breakdown above the form.
Use Download CSV or Download PDF to share results.

Practical workflow: start with budget, then iterate sample rate, indexed percent, and retention until total cost fits.

Sampling Strategy and Signal Coverage

Sampling rate is the fastest lever for controlling trace spend. If your services produce 250 traces per second and you sample 20%, you keep 50 traces per second for analysis. Across a 30‑day month, that becomes roughly 131 million sampled traces. Use higher sampling only on critical paths, error responses, or latency outliers to increase signal without capturing everything, and review sampling rules after major traffic shifts.

Span Volume Drivers You Can Measure

Span volume is the next driver, because ingestion is priced per span and storage grows with span bytes. Multiply sampled traces by average spans per trace to estimate sampled spans per month. A typical web request might create 15-30 spans; deeper microservice chains can exceed 60. Trim high‑cardinality attributes, reduce large event payloads, and standardize span size assumptions using real exporter metrics to prevent cost surprises when teams add instrumentation.

Retention Policy and Storage Curve

Retention determines how long data occupies storage, so costs scale with days kept. The calculator estimates average stored gigabytes as daily stored GB multiplied by retention days, which approximates a rolling window in steady state. If you ingest 300 GB per month after compression, that is about 10 GB per day; at 14 days retention, average storage is near 138 GB. Align retention with incident response needs and compliance rules.

Indexing and Query Efficiency Tradeoffs

Indexing improves search and aggregates, but it adds extra storage. Treat indexed percent as the share of stored trace data that becomes searchable, then apply an overhead factor to represent index structures. For example, indexing 40% of stored traces with a one-and-a-half factor means index data is about 60% of stored trace GB. Keep indexes lean by indexing only query‑critical fields and using sampling for exploratory analysis.

Budget Guardrails and Scenario Reviews

Budget control works best with repeatable scenarios. Start with a target monthly ceiling, then test conservative, balanced, and deep‑dive configurations. Track unit economics like cost per 1000 sampled traces and cost per 1M sampled spans to compare services fairly. Add egress estimates if dashboards export large datasets. Revisit assumptions quarterly, or when traffic, span counts, or retention policies change materially.

FAQs

What does the sample rate represent?

It is the percent of incoming traces you keep. A 20% rate keeps about 1 in 5 traces, reducing ingestion and storage while still showing representative performance patterns.

How can I estimate average spans per trace?

Use exporter metrics or a trace backend report to compute total spans divided by total traces over a typical day. Start with 15–30 for simple web flows and adjust after adding downstream instrumentation.

Why can retention increase storage cost quickly?

Storage scales with the retention window. If daily stored volume is 10 GB, moving from 14 to 30 days increases average stored GB from about 140 to 300, before indexing overhead.

When should I include indexing in the estimate?

Include it if you rely on fast search, tag filtering, or high‑cardinality queries. Keep indexed percent focused on essential fields and validate the overhead factor using your vendor’s documented index behavior.

How do span size and compression affect results?

Span size drives raw bytes, and compression converts raw to stored size. Large attributes, events, and logs inflate spans. Reducing payload and using consistent attribute conventions usually cuts both storage and egress.

Which pricing inputs should come from my provider?

Update ingestion price per million spans, storage per GB‑month for traces and indexes, and network egress per GB. If you pay a base platform fee, add it as a fixed monthly amount.