| Scenario | Logs GB/day | Metrics M/day | Traces GB/day | Seats | Monthly estimate |
|---|---|---|---|---|---|
| Startup (prod + staging) | 25 | 40 | 8 | 15 | $1,150 |
| Mid-market SaaS | 120 | 180 | 35 | 60 | $5,980 |
| Enterprise (multi-region) | 600 | 900 | 220 | 250 | $32,400 |
This calculator uses a transparent, vendor-neutral model:
- LogsEffectiveGB/day = LogsGB/day × (1 − Compression%)
- TracesEffectiveGB/day = TracesGB/day × (SamplingKept%)
- MonthlyIngest = DailyVolume × 30
- IngestCost = MonthlyIngest × UnitPrice
- StoredGB ~= EffectiveDailyVolume × RetentionDays
- StorageCost = StoredGB × StoragePricePerGBMonth
- SeatCost = Users × PricePerUserMonth
- CheckCost = Checks(M) × PricePerMillionChecks
- Subtotal = Sum(all line items)
- Total = (Subtotal + Overhead%) − Discount%
- Start with today’s measured daily volumes for logs, metrics, and traces.
- Set realistic retention by environment (prod vs non-prod).
- Apply compression and sampling based on pipeline settings.
- Enter unit prices from your preferred observability provider plan.
- Add seats, alert checks, and optional APM or synthetic monitors.
- Use overhead and discount for support, add-ons, or negotiations.
- Submit, review the breakdown, then export CSV or PDF.
Ingest volume is the first budget lever
Monthly ingest is estimated with a 30-day multiplier, so small daily changes compound quickly. For example, reducing logs from 120 GB/day to 90 GB/day cuts monthly ingest by 900 GB. At $0.35 per GB, that single change reduces ingest cost by $315 per month, before storage effects.
Retention converts flow into stored footprint
Retention is modeled as StoredGB ≈ EffectiveDailyVolume × RetentionDays, which mirrors steady-state behavior in a stable workload. If effective logs are 60 GB/day and retention is 30 days, the searchable store trends toward 1,800 GB. At $0.06 per GB-month, that is $108 per month for storage alone.
Sampling and parsing shape trace economics
Trace sampling applies directly to trace volume. Keeping 25% instead of 50% halves trace ingest and trace storage, which is useful when high-cardinality spans dominate. For logs, structured parsing and field filtering often deliver 15–45% reduction; the calculator treats this as a single compression percentage to simplify planning.
Seats, checks, and add-ons become dominant at scale
As organizations grow, seat licensing and alert checks can exceed data charges. A team of 250 users at $18 per user costs $4,500 monthly regardless of ingest. Similarly, 300 million alert checks at $0.10 per million adds $30 per month, but higher-priced checks or complex queries can shift this quickly.
Use scenarios to negotiate and to govern usage
Run three scenarios: conservative (lower volumes), expected (current measurements), and peak (incident or launch). Compare the subtotal, then apply overhead and discount to reflect support tiers, SSO requirements, or annual commitments. The resulting breakdown supports chargeback by service, retention policies by environment, and sampling targets by criticality.
FAQs
1) Does this model match every vendor’s billing?
No. It is a vendor-neutral estimator that mirrors common levers: ingest, retention, seats, checks, and add-ons. Map your plan’s units to the closest input fields for a reliable forecast.
2) Why is a 30-day month used?
Using 30 days keeps results predictable for planning. If you need calendar-accurate totals, rerun the calculator with adjusted daily volumes or update the month multiplier in the code.
3) How should I choose the metrics storage factor?
Start with 0.05–0.15 GB per million retained samples, then calibrate using actual backend storage reports. High-cardinality labels and longer series increase the factor over time.
4) Where do compression and sampling percentages come from?
Use pipeline settings and historical data. For logs, compare raw vs indexed volume. For traces, compare spans received vs stored. Update the percentages as telemetry governance matures.
5) What overhead should be included?
Overhead can represent premium support, enterprise SSO, compliance features, data transfer, and taxes. Keep it modest and transparent, then adjust once real invoices confirm the variance.
6) Can I export results without server libraries?
Yes. CSV and PDF downloads are generated in the browser using the computed results. This keeps deployment simple while still producing shareable artifacts for finance reviews.