Safety Over Liveness: Breaking Down the Uptime Metric for Validator Performance

Key points

• Uptime is an over-simplified metric that doesn’t accurately capture performance
• Slashing risks are more material than minor downtime / inactivity
• Network downtime tolerance could mean immaterial penalties for inactivity
• Network reward structures could mean full rewards despite inactivity
• Figment provides Slashing, Downtime, and Missed Rewards coverage, ensuring real rewards protections rather than relying on uptime guarantees

Introduction

Not all metrics are created equal when attempting to compare and contrast staking providers on a Proof-of-Stake (PoS) or Delegated Proof-of-Stake (dPoS) network. By only measuring how often a validator is online, uptime is typically a vanity metric as it ignores a validator’s actual performance. If fixated on by a staking provider, uptime could adversely affect its delegators. A better set of metrics focus on block production and safe signing practices.

At Figment, we advocate for a “ safety over liveness” approach, coupled with comprehensive Slashing, Downtime, and Missed Rewards coverage. This strategy offers our customers a more effective way to maximize risk-adjusted rewards and protect their staked assets rather than relying on abstract uptime guarantees.

Through examples from Solana, Ethereum, and Cosmos, we’ll demonstrate why fixating on uptime can be misleading and potentially harmful to stakers.

The Uptime Myth: Why 99% Online Doesn’t Equal 99% Performance

Uptime is an overly simplified metric that does not measure block production quality. Since uptime only measures whether a server is online, it tends to over-simplify the situation  and not give the full picture of a validator’s performance. A validator can have high uptime but still fail to perform its duties effectively. A more insightful measure of validator health would consider participation factors like timely block production and block validation as these correlate to rewards earned. An alternative metric like skip rate measures how often a validator “skips” their turn to produce a block. For instance, a validator could be online and therefore said to be achieving nearly 100% uptime, while failing to propose or attest to blocks correctly. Despite meeting the 100% uptime metric, this same validator would likely earn below average rewards or even incur penalties.

Solana Uptime Example:

• On Solana you will find many validators have an uptime of greater than 99%, but no validators have a slot success rate of more than 99% (calculated as 100% less Skip Rate). This would imply that a validator with a 99% uptime guarantee would meet its minimum threshold, but could be earning below average rewards.
• In the example below, the anonymous validator has uptime of 99.75%, but its skip rate is 13.18%. The resulting slot success rate is only 86.57%. Rewards earned staking on this validator would be lower over time than a validator with a comparatively lower uptime percentage of say 95%, but a comparatively better skip rate over 90%.

The Slashing Risk: Why Safety Can Beat Liveness

Slashing risks are more material than minor downtime. Many PoS and dPoS networks utilize slashing penalties as a mechanism to maintain security and ensure validators act in the best interest of the network. Misbehaving validators can incur severe penalties which can result in the loss of a portion or all of their staked tokens. The achilles heel of a high uptime metric is that it does not measure slashing performance nor does it protect against it. Delegators need to look at both a validator’s slashing history and consider whether a validator follows the best security practices, abides by the protocol, and avoids risky behaviors.

Ethereum Uptime Example:

• On Ethereum, a 99% uptime guarantee would imply a tolerance of ~7.2 hours of downtime per month. 7.2 hours of downtime translates to roughly 0.0005ETH in penalties per validator, or $1.29 USD at the time of writing. Estimated missed rewards during that same 7.2 hour window would be roughly equivalent totaling $2.58 in penalties and missed rewards
• If this validator was to be double-sign slashed, it would lose 1 ETH of its effective balance immediately, followed by ~0.05ETH over a 36-day period. At the time of writing, this equates to a $2,703.77 USD slashing penalty. Measured in uptime, this validator would need to be down 15,090 hours to incur this same penalty
• Proper validator practices and protection against slashing far outweigh what any 99% uptime guarantee can protect

The 5% Paradox: When 95% Downtime Doesn’t Cost You

Network downtime tolerance could mean no penalties for extensive inactivity. PoS and dPoS networks have varying tolerance for validator downtime. For many networks, if a validator has a brief downtime but quickly recovers without missing crucial blocks, it may not impact rewards significantly. Focusing purely on a validator’s uptime, without understanding the network’s specific penalty structure, can lead to poor decision-making. Validators with slightly lower uptime may still offer better long-term rewards, particularly if they exhibit resilience and handle network conditions effectively.

Cosmos Uptime Example:

On Cosmos, a validator only incurs downtime penalties should it miss 9,500 out of 10,000 blocks. This implies that a validator could be offline for a 19 hour period over a 20 hours window and not incur any downtime penalties. An uptime guarantee for Cosmos would not make sense as there would not be penalties incurred so long as the validator is online 5% of the time.

Full Rewards for 5% Uptime: Breaking Down the Network Incentives

Network reward structures could mean full rewards despite inactivity. Like network-specific downtime tolerances, PoS and dPoS networks have varying reward structures that are impacted differently by downtime. Many networks will reward validators with the maximum amount of inflationary rewards, so long as it is online and performing its duties above a specified threshold.

5% Uptime Network Example:

On Cosmos, a validator would earn 100% of inflationary rewards, so long as it did not miss more than 95% of the last 10,000 blocks. This means that a validator can be offline for roughly 19 hours out of every 20 hours and still earn nearly all inflationary rewards as though it had been up for 100% of the time. An uptime guarantee for Cosmos would not make sense as rewards would hardly be impacted so long as the validator is online 5% of the time.

Conclusion

There is no such thing as a “one-size-fits-all” uptime metric that could be applied across networks. In fact, uptime percentages should be seen as a vanity metric, with little to no impact on validator performance.

Network Uptime Examples: Key points

• Solana: The skip rate is a better proxy for rewards earned for delegators.
• Ethereum: Prioritizing uptime could result in far more material slashing penalties.
• Cosmos: Extensive periods of inactivity could result in no downtime penalties incurred and the same amount of inflationary rewards being earned.

Figment firmly believes in its safety over liveness operating principle as the best strategy for its customers to earn the highest risk-adjusted rewards. This approach is validated through our consistent rewards performance on 30+ networks. Those who have staking agreements with Figment get the benefit of Figment’s Slashing, Downtime and Missed Rewards coverage; a more relevant approach to protecting underlying staked assets and foregone rewards due to extensive downtime. Figment demonstrates that a security-focused strategy ultimately delivers better risk-adjusted rewards for our customers.