Observability platforms, compared
Each provider is profiled from its own documentation, with the source recorded against every figure. What a plan includes is read from whichever meter that provider actually publishes — events, or a price per gigabyte ingested — since most document one and not the other.
Error tracking and observability
| Provider | What a plan includes | Retention window |
|---|---|---|
| Axiom | Not counted in events: 500 GB/month data loading, 10 GB-hours query compute and 25 GB storage on free Personal; 1 TB, 100 GB-hours and 100 GB included in the Cloud platform fee. | 30 days on the free Personal plan; configurable on Cloud rather than fixed. |
| Better Stack | 100,000 exceptions and 5,000 session replays a month on the free personal plan, plus 30 GB metrics and 3 GB each of logs, traces and web events. Six meters, priced separately: exceptions at $0.000050 each, replays at $0.00150, metrics at $0.50 per GB per month. | 90 days for errors and session replays. Logs, traces and web events are retained 3 days on the free personal plan, where retention is what the 3 GB allowance buys rather than a fixed window. |
| Datadog | None included. Log indexing is $1.70 per million events per month at 15-day retention, and Flex Storage is $0.05 per million events stored — ingest and indexing are separate charges on the same logs. | Chosen per index and priced with it: 3, 7, 15 or 30 days, or greater than 30 by arrangement. The $1.70 per million rate is the 15-day tier. Infrastructure metrics retain 1 day on the free plan. |
| Elastic Cloud | As low as $0.07 per GB ingested (logs, traces, RUM) on Serverless Observability Logs Essentials; as low as $0.09 on Observability Complete, with metrics at $0.023. | No fixed window on Serverless: retention is priced per GB retained per month, as low as $0.017 (Logs Essentials) or $0.019 (Complete), and $0.005 for metrics. |
| GlitchTip | 100,000 events a month on the cheapest paid plan; 1,000 on the free hosted plan. One meter, not several — GlitchTip counts events and nothing else. | 90 days by default when self-hosted, set by GLITCHTIP_RETENTION_DAYS and configurable per data type — events, transactions, logs, uptime checks and files all inherit it, releases keep 365 days. Events move from Postgres to cold storage after 30 days but stay queryable. GlitchTip publishes no retention figure for its hosted plans. |
| Grafana Cloud | Free plan, per month: 10k active metrics series; 50 GB each of logs, traces and profiles; 50k frontend-observability sessions; 500 k6 virtual user hours; 100k API test executions. Grafana meters series, GB, sessions and host hours; no event-count meter is published. | 13 months for metrics and 30 days for logs, traces, profiles and k6 tests on Pro. The free tier is 14 days for everything. |
| OpenObserve | $0.50 per GB ingested, quoted with a 30% annual-commitment discount already applied. | Included rather than priced: 15 months for metrics and 30 days for logs, traces, RUM and everything else, at no extra cost on pay-as-you-go. |
| Sentry | 50,000 errors a month on Team and Business; 5,000 on Free. Errors are one meter of several — logs, spans, replays and uptime monitors are counted and billed separately. | 30-day lookback on Developer, the free plan. Up to 90 days on Team, and up to 90 days plus additional sampled retention on Business. Enterprise is custom. "Up to" is Sentry's own wording — the 90 days is a ceiling, not a guarantee. |
| SigNoz | No events meter: SigNoz Cloud bills traces and logs per GB and metrics per million samples. The $49/month Teams minimum includes $49 of usage — e.g. 163 GB of logs or traces, or 490 mn metric samples. | 15 days by default for logs and traces, selectable up to 30, 90, 180 days or 1 year at higher rates. |
Which to consider
Small projects that want a real telemetry platform for nothing — 500 GB a month loaded with 30-day retention, permanently, is the most generous free allowance in this category.
Teams who would otherwise buy uptime monitoring, log management and error tracking separately, and would rather have one bill and one dashboard.
Teams that need one vendor across infrastructure, logs, traces and replay, and can model a bill that has a separate meter for each.
Teams that need full-text search and observability on the same underlying engine, or that want the option to self-host later — Elastic's dual identity as a search platform and an observability platform is the actual differentiator against single-purpose competitors.
Anyone who wants Sentry-shaped error tracking without Sentry's bill — the SDKs are compatible, so switching is a DSN change, and self-hosting is free and genuinely supported.
Metrics-heavy monitoring where 13 months of retention matters — that window is an order of magnitude beyond what the error trackers and most telemetry platforms keep.
Teams that want one bill they can predict — ingest, query, and retention included — with session replay and RUM in the same product rather than as separate purchases.
Teams that want tracing, logs, replays and uptime in one place and will pay per meter for it. The tooling is the deepest in the category and the SDK coverage is the widest.
Teams that want per-GB billing without per-host or per-seat charges, and the option to run the same product themselves rather than being locked to the cloud version.
Observability in the news
- Rest-Of-Latin-America Observability Tools & Platforms Market Size, Share,Trends, Growth Analysis Report, 2031 - MarketsandMarkets
- The Observability Tax: When Monitoring Costs Exceed Downtime Costs - DevOps.com
- South-Africa Observability Tools & Platforms Market Size, Share,Trends, Growth Analysis Report, 2031 - MarketsandMarkets
- The Best OpenRouter Alternatives in 2026 - Technology Org
Comments are hosted by EchoThread, which sets cookies and offers social sign-in. Like the analytics on this site, it does not load until you accept — see the privacy page.