TL;DR: For error tracking, Sentry gives 5,000 errors and 5 million spans a month on one seat, while Bugsnag trades 50% more monthly events for a fifth of the span budget and a week of retention. GlitchTip and Bugsink are the smallest and simplest, built to self-host if you outgrow the hosted quota. For full observability, Grafana Cloud is the widest free tier here - 10,000 metrics series plus 50 GB each of logs, traces, and profiles a month, no card required - and New Relic backs it with 100 GB of ingest on one full platform seat. Datadog’s free Infrastructure tier is the tightest of the group at 5 hosts and a single day of retention. For uptime and status pages, UptimeRobot covers 50 monitors at 5-minute intervals, and Healthchecks.io watches 20 cron jobs for free. Full comparison and exclusions below.
What’s in this list, and what isn’t
This covers three related jobs: catching and triaging application errors, watching infrastructure and request traces end to end, and confirming a service or a scheduled job is still alive. A few products that publish something on their pricing page don’t make the tables:
- highlight.io is out. Its pricing page and its marketing site both now 308-redirect to LaunchDarkly, with no independent product or free-tier page left to check a quota against.
- APItoolkit is out under its current name, Monoscope. Its live pricing page shows a paid Cloud plan starting at $29/month and a self-hosted “Community Edition” that’s free forever, but no managed, hosted free tier matching what a request-tracing free plan on this list needs.
- Statuspile is out because its site publishes no plan, price, or usage limit at all, so there’s no free-tier figure to compare. The open-source project behind it aggregates other vendors’ status pages rather than watching your own services, a different job from the rest of this table.
- OpenObserve’s free tier is the self-hosted Enterprise edition, free for up to 50 GB of ingestion a day - not the hosted Cloud plan, which is a 14-day trial. It’s in the table under that reading, not the trial one.
Everything below is checked against the vendor’s own pricing page as of September 2026.
Error tracking and crash reporting
| Product | Free tier | Gotcha |
|---|---|---|
| Sentry | 5,000 errors/month, 5M spans/month, 50 session replays/month, 1 user | 30-day retention, 1GB attachment storage |
| Bugsnag | 7,500 events/month, 1M spans/month, 1 user | 7-day data retention |
| Rollbar | 5,000 occurrences/month, 1,000 sessions/month | 30-day retention cap - 90 days starts on Essentials |
| Honeybadger | 5,000 errors/month, 1 uptime monitor, 1 status page, 1 user | 15-day error retention, 7-day log retention |
| GlitchTip | 1,000 events/month, unlimited projects and team members | Uptime checks, performance transactions and release uploads all draw on that same event quota |
| Bugsink | 15,000 monthly events, 5,000 events retained | Single user; built to self-host past this |
| Bugfender | 100,000 log lines/day, 2,000 devices | 24-hour log retention |
| Embrace | 1M sessions/year, 5 users | Only 3 days of event retention |
| PostHog | 100,000 exceptions/month for error tracking | Error tracking is one product inside a much larger analytics suite |
| LogDog | 100 MB of data/month, forever | Built for mobile device logs, not server errors |
| Shake | 20 bug reports/month, 3 team members | Unlimited crash reports, but bug-report volume is the real cap |
Sentry is the default choice for a reason: 5,000 errors and 5 million spans a month cover a small production app comfortably, and the same plan throws in 50 session replays, an uptime monitor, and a cron monitor at no extra cost. One seat is the hard limit, so a two-person team hits the wall on users before they hit it on events. Bugsnag is the closest like-for-like alternative, and the trade runs in both directions: 7,500 events a month against Sentry’s 5,000, but 1 million spans against Sentry’s 5 million, on the same single-seat ceiling. Its 7-day retention window is a quarter of Sentry’s 30 days, so a bug that surfaces after a week is already gone from the dashboard.
Rollbar and Honeybadger both land at 5,000 events a month, and both are single-user free plans. Honeybadger folds in a free uptime monitor and status page most error trackers charge extra for, but caps log and Insights data at a week of retention. GlitchTip and Bugsink are the two options built around self-hosting from day one: GlitchTip’s hosted free plan tops out at 1,000 events a month, and “event” is broader than it sounds: error occurrences, uptime checks, performance transactions, and release file uploads all draw on the same budget, so turning on performance monitoring eats into the errors you can afford to record. Bugsink only retains 5,000 of its 15,000 monthly events, on the assumption that a project outgrowing that runs its own instance instead of paying more to the vendor.
Bugfender and Embrace both track mobile apps specifically rather than general server errors. Bugfender’s free plan is sized by log volume - 100,000 lines a day across 2,000 devices - with logs disappearing after 24 hours. Embrace measures in sessions instead: 1 million a year across up to 5 team members, but with only 3 days to look back at what went wrong. PostHog earns a place despite error tracking being a small slice of what it does: 100,000 exceptions a month is a real, permanent allowance that renews alongside its analytics, session replay, and feature flag quotas. Shake and LogDog are the smallest entries - Shake caps at 20 bug reports a month though crash reports stay unlimited, and LogDog is a mobile log viewer with a 100 MB/month ceiling rather than a full error tracker.
Full observability, APM, and logs
| Product | Free tier | Gotcha |
|---|---|---|
| New Relic | 100 GB of data ingest/month, 1 full platform user | Unlimited basic users, but only one full user with all 50+ capabilities |
| Grafana Cloud | 10,000 metrics series, 50GB logs, 50GB traces, 50GB profiles/month | 14-day retention across the board, 3 active users |
| Honeycomb | 20M events/month, 100M metrics data points/month | Requires structured event data, not a drop-in log tail |
| Axiom | 500 GB of data ingest/month | 30-day retention |
| Datadog | 5 hosts monitored | 1-day metric retention on the free Infrastructure tier |
| Amazon CloudWatch | 5 GB logs ingested/month, 10 custom metrics, 10 alarms | Requires an AWS account; alarms only cover standard-resolution metrics |
| Better Stack | 10 monitors, 100,000 exceptions/month | Logs and traces both capped at 3 GB, 3-day retention |
| Inspector | 30,000 transactions/month | Up to 3 team members |
| Skylight | 100,000 traces/month on any account | Free without limit for open-source apps specifically |
| OpenObserve | Self-hosted Enterprise free for up to 50 GB ingestion/day | Requires running the software yourself; the hosted Cloud plan is a 14-day trial |
Grafana Cloud covers the most ground of any product on this list: 10,000 active metrics series plus 50 GB a month each of logs, traces, and profiles, with no credit card and no expiry date. The tradeoff is retention - everything drops off after 14 days, and only 3 people can actively use the account. New Relic takes a different shape: 100 GB of ingest a month sounds smaller than Grafana’s combined allowance but comes with unlimited basic users for querying dashboards and alerts, gated behind a single full-access seat with all 50-plus platform capabilities.
Honeycomb and Axiom both lean into raw volume: Honeycomb’s 20 million events a month is built around structured, high-cardinality event data rather than plain log lines, while Axiom’s 500 GB ingest a month with 30-day retention reads more like a general-purpose log store. Amazon CloudWatch is the smallest of the AWS-native options at 5 GB of log ingestion and 10 custom metrics a month, useful mainly if the rest of a stack already runs on AWS and an AWS account is a given anyway.
Datadog’s free Infrastructure tier is the tightest one here: 5 hosts and a single day of metric retention, enough to evaluate the product but not to run production monitoring on. Better Stack and Inspector both sit in the middle - Better Stack bundles 10 uptime monitors with 100,000 tracked exceptions a month, though logs and traces are limited to 3 GB with a 3-day window; Inspector gives 30,000 transactions a month described by the vendor as a standing quota rather than a trial, capped at 3 team members. Skylight splits its offer in two: 100,000 traces a month for any account, unlimited for open-source projects specifically. OpenObserve’s real free tier is the Self-Hosted Enterprise edition - free up to 50 GB of ingestion a day, with SSO, RBAC, and audit trail, if the team runs the software itself. The hosted cloud version is a 14-day trial.
Uptime, cron monitoring, and status pages
| Product | Free tier | Gotcha |
|---|---|---|
| UptimeRobot | 50 monitors | 5-minute check interval on the free plan |
| Healthchecks.io | 20 monitored jobs, 100 log entries/job | Same limits on the paid Supporter tier |
| Cronitor | 5 monitors | Fastest check frequency is 5 minutes - 30 seconds starts on the paid plan |
| StatusGator | 3 monitors, 1 board, 100 subscribers | 1 month of historical data |
| UptimeToolbox | 5 monitors, 1 server | 3-minute check interval |
| MonitorMonk | 2 dashboards, 10 websites, SSL and keyword checks | An SLA and premium support start on the paid plans |
| PingPong | Unlimited incidents, components, and scheduled maintenance; 60-second metric checks | Team seats start on Pro - 10 there, 25 on Business; Solo has none |
UptimeRobot covers the most monitors of anything in this group - 50 - checked every 5 minutes, which is enough for most side projects and small production sites. Healthchecks.io solves a narrower problem well: it watches 20 scheduled jobs rather than websites, alerting when a cron job goes quiet instead of when a URL stops responding, and the free plan’s limits don’t tighten if a project later pays for the Supporter tier - the money buys goodwill, not headroom. Cronitor does both jobs and URLs on one free plan but at a fifth of UptimeRobot’s monitor count.
StatusGator, UptimeToolbox, and MonitorMonk are built around a status page rather than a monitor fleet - StatusGator watches 3rd-party status pages across 3 monitors, UptimeToolbox pairs 5 monitors with 1 status page on a 3-minute interval, and MonitorMonk keeps it simplest at 2 dashboards and 10 websites, with an SLA reserved for the paid plans. PingPong stands out for capping nothing that matters on a status page - incidents, components, and scheduled maintenance windows are all unlimited on the free plan, and its metric checks run every 60 seconds, the fastest interval in this group. Named team seats are the thing it holds back: they start on Pro at 10, not on the free plan.
Which one should you actually use?
- A single-seat app that needs real error detail: Sentry for the biggest span budget and 30-day retention on one free seat; Bugsnag if you’d rather have 50% more monthly events and can live with fewer spans and a shorter window.
- A team that wants a free uptime monitor and status page bundled with error tracking: Honeybadger.
- The widest free observability stack, if 14-day retention is enough: Grafana Cloud - metrics, logs, traces, and profiles all in one free plan.
- One full-featured APM seat with a bigger data budget: New Relic at 100 GB a month.
- Just confirming a service or a job is still alive: UptimeRobot for the most monitors, Healthchecks.io for cron jobs specifically.
- Planning to self-host once you outgrow a hosted free tier: GlitchTip or Bugsink - both self-host with unlimited users.
Whichever you pick, check the product page for the exact numbers before you build against it - free tiers change, and a retention window matters more day to day than the headline quota.
FAQ
Is Sentry actually free forever? Yes. The Developer plan is a permanent free tier: 1 user, 5,000 errors a month, 5 million spans, and 50 session replays, with no expiry date and no credit card charge if usage stays inside those limits.
What’s the most generous free observability platform? Grafana Cloud covers the widest range of signal types free - 10,000 metrics series plus 50 GB each of logs, traces, and profiles a month - though everything is capped at 14 days of retention and 3 active users.
Is there a free uptime monitor that doesn’t require a credit card? UptimeRobot and Healthchecks.io both run permanent free plans with no card required - 50 monitors on UptimeRobot, 20 monitored jobs on Healthchecks.io.
What’s the smallest free tier in this list, if I just want to try something out? GlitchTip at 1,000 events a month and Cronitor at 5 monitors are both built for evaluation-scale usage before a self-hosted or paid upgrade.
Is Datadog worth using for free? Only for evaluation. 5 hosts and a single day of metric retention is enough to see how the product works, but not enough to run ongoing production monitoring without upgrading.
See also
- The Best Free Cloud Hosting in 2026 - where the apps generating these errors and traces actually run
- Free Background Jobs and Cron in 2026 - scheduled jobs and queues, plus Healthchecks.io and Cronitor covered from the job-monitoring angle
- The Best Free Database Tiers in 2026 - free-tier storage for the state your monitored services read and write
- Error tracking, observability platforms, APM tracing, log management, uptime monitoring, and status pages - every free-tier product in each category, filterable
- FTV methodology - how the dollar value on each product page is estimated