Observe pricing guide details
Pricing Overview
Observe (now Observe by Snowflake) prices by the data you send, not by host or seat. There are three published rates, one per signal type:
- Logs: $0.49 per GiB
- Traces: $0.59 per GiB
- Metrics: $0.008 per data point per month (written as /DPM on the pricing page)
Each rate includes compute, unlimited users and standard retention: 30 days for logs and traces, 13 months for metrics. Extended retention is $0.01 per GiB per month. Observe sells this as a subscription based on a committed volume of uncompressed telemetry data ingested, and states that multi-year and volume-based discounts are available. A free trial is offered without a credit card.
The model suits teams with many engineers and lots of hosts, because neither adds a fee by itself. What you pay for is volume, so a team that controls what it sends controls its bill.
Plan Comparison
Observe has no named plans. It has one rate card, split by signal:
| Signal | Price | Compute | Users | Standard retention |
|---|---|---|---|---|
| Logs | $0.49 per GiB | Included | Unlimited | 30 days |
| Traces | $0.59 per GiB | Included | Unlimited | 30 days |
| Metrics | $0.008 per data point per month | Included | Unlimited | 13 months |
| Extended retention | $0.01 per GiB per month | — | — | Beyond the standard period |
The key point is that metrics are not billed by volume. Logs and traces are counted in GiB, but metrics are counted in data points. Observe's pricing page does not define how a data point is counted for a given scrape interval, so ask before you estimate a metrics-heavy workload. That count, not the size of the data, drives the metrics line.
Traces cost more per GiB than logs. A team that sends full traces for every request will see that line grow faster than its log line.
Hidden Costs and Considerations
- You commit to a volume. Pricing is based on a committed volume of uncompressed data. Observe says it never sends an overage bill and will work with you to resize capacity if you go over, but the commitment itself is what you pay for, so size it with care.
- Uncompressed data counts. The rate applies to uncompressed telemetry, so compression on your side does not shrink the bill.
- Long retention is cheap but adds up. At $0.01 per GiB per month, keeping 10,000 GiB of logs beyond the standard 30 days costs 10,000 x $0.01 = $100 for each extra month.
- Discounts are by quote. Multi-year and volume discounts exist, but no discount rates are published.
Cost Estimates by Team Size
Observe charges nothing per user, so team size does not change the bill. Data volume does. The scenarios below state the workload and apply the published rates:
- 150 GiB of logs a month: 150 x $0.49 = $73.50 per month.
- 1,500 GiB of logs and 500 GiB of traces a month: 1,500 x $0.49 + 500 x $0.59 = $1,030 per month.
- 6,000 GiB of logs and 2,000 GiB of traces a month: 6,000 x $0.49 + 2,000 x $0.59 = $4,120 per month. Metrics come on top at $0.008 per data point per month. Get Observe's data-point count for your own metrics before adding that line, since it depends on how Observe counts your series. Whether 5 or 50 engineers use the platform, these figures stay the same.
How Observe Pricing Compares
All competitor prices below come from each vendor's current pricing record.
- Datadog charges per host for infrastructure, from $15 per host per month billed annually on Pro. Logs are priced separately: $0.10 per GB per month for ingestion, plus $1.70 per million indexed logs per month billed annually, with 15-day retention.
- New Relic includes 100 GB of data ingest per month free, then $0.40 per GB (original) or $0.60 per GB (Data Plus). It also charges for users: Pro is $349 per full platform user per month on annual commitment.
- Grafana Cloud Pro is from $19/month plus usage, with logs and traces starting at $0.050/GB process, $0.400/GB write and $0.100/GB retain.
Observe's per-GiB rate looks higher than some of these ingest rates, but it includes compute, 30 days of retention and every user. Compare the full bill for your own data mix, not the headline rate.