Why is my crash-free rate dropping?

Because crashes are increasing, usually from a new crash a recent release introduced or an existing crash surging, the drop is driven by specific crashes that per-version data and impact ranking can pinpoint. Crash-free rate (the share of users or sessions that don't experience a crash) drops when crashes increase, so a falling rate means more of your players are crashing, and there's a specific cause. The most common: a recent release introduced a new crash (a regression in an update that crashes players, increasing the crash count and dropping the rate, the rate falls as players hit the new crash), an existing crash surged (a crash that became more frequent, perhaps due to more players hitting its conditions, a content change, or an external factor like an OS update), an OS update broke something (players' devices updating and the new OS version crashing your game), or a growing/changing player base hitting crashes (new devices or conditions surfacing crashes). The most common by far is a release-introduced crash, since releases are when new crashes enter. To find the cause, check per-version data (if the rate dropped on a specific build, that release introduced the driving crashes) and identify the specific crashes (rank by impact, find the new or surging signatures that appeared as the rate dropped). A crash-free rate drop is almost always driven by one or a few specific crashes increasing, not a general decline, so pinpointing them is the key. So your crash-free rate is dropping because crashes are increasing, usually from a new crash a recent release introduced (or an existing one surging), which per-version data and impact ranking pinpoint. Bugnet tracks crash-free rate and crashes per version with impact ranking, so when your rate drops you can see which build it started on (the cause, usually a release) and which crashes are driving it (the new or surging signatures), turning a falling metric into a specific diagnosis, so you can fix the driving crashes (or roll back the release) and verify the crash-free rate recovers.

What is a good crash-free rate for a game?

It depends on platform and game, but generally you want crash-free users/sessions as high as possible, often 99%+ is a common target, the key is tracking your rate over time and reacting to drops, more than hitting a universal number. Crash-free rate (typically measured as crash-free users, the percentage of users who don't experience a crash in a period, or crash-free sessions, the percentage of sessions without a crash) is a key stability metric, and what's good depends on context. General benchmarks: many developers target 99%+ crash-free sessions or users as a healthy level, with higher (99.5%, 99.9%) being better, a rate well below 99% suggests crashes are affecting a meaningful share of players and there's significant room to improve. But the right target varies by platform (mobile, with its device fragmentation, may have different typical rates than PC or console), game type and complexity, and your player base's devices (a game on a wide range of low-end devices may face more crashes than one on a narrow high-end range). So rather than fixating on a universal number, the more useful approach is: track your crash-free rate over time and per version (knowing your baseline and watching for changes), aim to improve it (driving it higher by fixing high-impact crashes), and react to drops (a falling rate signals new crashes to address, regardless of the absolute number). A stable, high crash-free rate that holds across releases indicates good stability, while a dropping or low one signals problems. The trend and your ability to maintain/improve the rate matter as much as the absolute value. Also consider that crash-free rate alone doesn't capture everything (severity matters, a rare but game-breaking crash, freezes that aren't counted as crashes), so use it alongside other signals. So a good crash-free rate is generally as high as possible (often 99%+ as a common target), but the key is tracking your rate over time and per version and reacting to drops, more than hitting a universal number. Bugnet tracks your crash-free rate per version, so you can know your baseline, watch it over time and across releases, see when it drops (and which release or crash caused the drop), and confirm it recovers as you fix issues, helping you maintain and improve your crash-free rate by monitoring it and reacting to changes, which matters more than any single benchmark number.

How do I improve my crash-free rate?

Fix the high-impact crashes pulling it down, ranked by affected users, and catch regressions per version so releases don't drop it, since crash-free rate reflects how many users crash, fixing the crashes that affect the most users raises it most. Improving your crash-free rate (raising the share of users/sessions without a crash) means reducing crashes, specifically the ones affecting the most users, since the rate is driven by how many users crash. First, capture and rank crashes: capture crashes from the field with impact ranking, so you see which crashes affect the most users, these are what's pulling your crash-free rate down most, since each user who crashes reduces the rate. Second, fix the high-impact crashes: fix the crashes affecting the most users first, because crash volume is concentrated (a few crashes affect most of the crashing users), so fixing the top crashes removes most of the crashing, raising the crash-free rate the most per fix, this is the efficient path to a higher rate. Use the captured context to fix root causes. Third, catch regressions per version: monitor crash-free rate per version and watch for releases that drop it (a release introducing crashes lowers the rate), so you catch and fix (or roll back) regressions fast, preventing releases from dragging the rate down, maintaining your gains. Fourth, verify improvements: track the crash-free rate per version as you fix crashes, confirming it rises (the metric recovering/improving as the crashes stop), so you know your fixes worked. Fifth, sustain it: keep monitoring and fixing, since new crashes (from updates, OS changes, new devices) will arise, and maintaining a high crash-free rate is ongoing. The principle is that crash-free rate is driven by how many users crash, so fixing the crashes affecting the most users (high-impact, by impact ranking) raises it most, while catching regressions keeps releases from dropping it. So you improve your crash-free rate by fixing the high-impact crashes pulling it down (ranked by affected users) and catching regressions per version. Bugnet captures crashes with impact ranking (so you fix the crashes affecting the most users, raising the rate most) and tracks crash-free rate per version (so you catch regressions that drop it and verify your fixes raise it), so you can systematically improve your crash-free rate by fixing what affects the most users and maintaining the rate across releases, with the per-version tracking confirming the rate rises as you fix.

What to Do When Your Crash-Free Rate Is Dropping

Quick answer: Find what's driving it by checking per-version data (a recent release often introduces the drop) and identifying the specific crashes that rose, fix the high-impact crashes pulling the rate down, and verify the crash-free rate recovers.

A dropping crash-free rate, the share of users or sessions without a crash, falling, means crashes are increasing. Finding what's driving the drop is the key to reversing it. Here is what to do when your crash-free rate is dropping.

Check Per-Version Data for What Changed

A dropping crash-free rate has a cause, often a recent release. Check per-version data: if the rate dropped on a specific build, that release introduced the crashes pulling it down. Per-version comparison isolates the drop to a version, pointing at what changed.

Bugnet tracks crash-free rate and crashes per version, so when the rate drops you can see which build it started on. That isolates the cause to a release (or rules one out), turning a falling metric into a specific lead, the build that introduced the crashes dragging the rate down.

Identify the Crashes Driving the Drop

Find the specific crashes pulling the rate down: rank crashes by impact and look for the new or surging ones that appeared as the rate dropped. A falling crash-free rate is usually driven by one or a few crashes increasing, identifying them gives you the exact thing to fix.

Bugnet ranks crashes by affected users and tracks them per version, so the crashes driving the crash-free rate drop, the new or surging signatures, are identifiable. With the specific high-impact crashes pulling the rate down and their captured context, you know exactly what to fix to restore the rate.

Fix the High-Impact Crashes and Verify Recovery

Fix the crashes driving the drop, by impact, or roll back the release that introduced them, then verify the crash-free rate recovers, rising back toward its prior level on the fixed build. Confirm the metric recovers, not just that you fixed something.

Bugnet tracks crash-free rate per version, so after fixing or rolling back you can watch the rate recover on the new build. This verifies the fix worked at the metric level, the crash-free rate climbing back as the driving crashes stop, confirming you addressed what was pulling it down rather than assuming.

When your crash-free rate is dropping, check per-version data for what changed (often a release), identify the specific crashes driving the drop by impact, fix them or roll back, and verify the rate recovers. A falling crash-free rate is usually driven by a new or surging crash you can pinpoint.