A cost calculator turns vendor pricing pages into a single comparable monthly number. Here is how to use our observability cost calculator well, and how to read what it gives you.
What to enter
| Input | What it means | Tip |
|---|---|---|
| Hosts | Hosts/nodes you monitor | Count Kubernetes nodes, not pods, for per-host vendors |
| Log GB / month | Total log volume ingested | Check a current bill or your log pipeline metrics |
| Retention (days) | How long logs stay hot/searchable | 7-15 days covers most active debugging |
How the estimate is built
The formula is deliberately transparent:
monthly estimate = (hosts x per-host rate) + (billable log GB x per-GB rate)
For each vendor with a published rate it applies that vendor’s numbers - Datadog $15/host Infra + $0.10/GB logs, Grafana Cloud ~$0.45/GB after 50 GB free, Elastic $0.105/GB, New Relic $0.40/GB after 100 GB free - and ranks them cheapest first. Known free allowances are subtracted automatically.
What it deliberately leaves out
To avoid fabricating numbers, the calculator excludes things that depend on your exact setup: APM/RUM/tracing add-ons, log indexing (Datadog’s $1.27/M events is often the biggest hidden cost), data transfer, custom-metric overage, support and discounts. Vendors that price per event (Sentry, Honeycomb), per vCPU (AppDynamics) or custom (Chronosphere) are listed as “not directly comparable” rather than guessed - because a fabricated number is worse than no number.
How to read the result
Treat the ranking as directional, not a quote. Use it to shortlist two or three vendors, then read their full pages and model the extras (indexing, RUM, retention) yourself. If you scale by hosts, watch the per-host vendors; if you scale by data, watch the per-GB ones. See per-host vs usage-based pricing for the trade-off.
Bottom line
A calculator is a shortlisting tool: input hosts + log GB + retention, compare the comparable vendors, then verify the extras on each vendor page. Prices are a June-2026 snapshot; verify on each vendor’s pricing page before relying on them. Try it now in the observability cost calculator.