Estimate log retention costs from daily volume, retention days, and per-GB rates. Compare self-hosted and SaaS logging providers.
Log retention costs can spiral out of control when daily ingestion volume is high and retention windows are long. SaaS logging platforms like Datadog, Splunk Cloud, and New Relic charge $0.10–$2.50 per GB ingested, and those costs multiply by every day of retention. Self-hosted solutions like Elasticsearch or Loki have lower per-GB costs but require infrastructure, maintenance, and operational expertise.
This calculator estimates the monthly cost of retaining logs based on daily volume, retention period, and the per-GB storage rate. It also provides a comparison column so you can evaluate two providers or deployment models side by side—for example, Datadog vs. self-hosted Elasticsearch, or different retention windows on the same provider.
By modeling costs before committing to a logging platform or retention policy, you can set realistic budgets and make data-driven decisions about which logs to keep, how long to keep them, and which provider offers the best value for your ingestion volume.
Logging costs are often the largest line item in observability budgets. This calculator helps you compare providers, optimize retention windows, and avoid bill shock by estimating costs before they hit your invoice. This quantitative approach replaces reactive troubleshooting with proactive monitoring, enabling engineering teams to maintain service level objectives and minimize unplanned system downtime.
stored_volume = daily_volume_GB × retention_days; monthly_cost = stored_volume × rate_per_GB_month; annual_cost = monthly_cost × 12
Result: $90/mo (Provider A) vs $15/mo (Provider B)
Stored volume: 10 GB/day × 30 days = 300 GB. Provider A at $0.30/GB = $90/month ($1,080/year). Provider B at $0.05/GB = $15/month ($180/year). Provider B saves $900/year. The tradeoff may be query speed, features, or operational burden.
When comparing logging providers, evaluate: per-GB ingestion cost, per-GB retention cost, query performance, alerting capabilities, correlation features, and operational burden. The cheapest provider is not always the best—fast query response during an incident can be worth the premium.
Drop DEBUG-level logs in production. Sample repetitive access logs (log 1 in 100 health checks). Use structured logging to avoid verbose stack traces in every log line. Aggregate counters into metrics instead of logging individual events.
Log volume typically grows 20–50% annually. Budget for next year's volume, not this year's. Set up alerts at 80% of your budget threshold. Negotiate rate cards with SaaS providers before you hit higher pricing tiers.
SaaS providers range from $0.10/GB (Grafana Cloud) to $2.50/GB (Splunk Cloud) for ingestion. Self-hosted Elasticsearch on cloud VMs costs roughly $0.03–$0.08/GB/month. On-premises solutions have hardware and labor costs instead.
Shorten retention for low-value logs. Filter out health check and debug logs before ingestion. Compress logs. Downsample old logs to keep only ERROR and WARN entries. Move old logs to cheap archive storage instead of deleting them.
It depends on the provider. Datadog bills on raw ingestion volume. AWS CloudWatch bills on ingested bytes before compression. Self-hosted Elasticsearch stores compressed data. Match your input volume to the provider's billing metric.
At scale (50+ GB/day), self-hosted is usually cheaper per GB. But factor in operational costs: Elasticsearch cluster management, upgrades, monitoring, on-call, and hardware refreshes. Below 10 GB/day, SaaS is often more cost-effective.
For debugging: 7–14 days of hot, searchable logs. For incident investigation: 30–90 days. For compliance: 1–7 years in archive storage. Set different retention per log level and service criticality to optimize cost.
Some providers charge an ingestion fee (per GB received) plus a separate storage fee (per GB retained). Others bundle both into a single rate. Clarify the billing model before comparing—a low ingestion fee with high storage rates can be more expensive overall.