2 ms·

> I guess my point is that if a rare bug is rare enough, then it doesn't really matter, especially if the effect isn't catastrophic. That's only true if you ru

by xorcist 11y ago

> I guess my point is that if a rare bug is rare enough, then it doesn't really matter, especially if the effect isn't catastrophic. That's only true if you run a trivial application. If you are a bank for example, it matters very much why the outcome of a transaction came out wrong, even if the effect wasn't catastrophic. Every circumstance and every bit of identity to the transaction is important. "It's rare" or "I don't believe in logging" doesn't cut it. > There are a lot more resources required to log everything all the time then to log a subset of things some of the time. Resources are a matter of budgetary concerns, mostly. What does a bug in production cost you? Are you liable? How fast? What's your SLA? Those things matter.

jedberg 11y ago
> Resources are a matter of budgetary concerns, mostly. True, but most people unfortunately don't have unlimited budgets to work with. Also, there is more than a monetary cost. Every system in your architecture incurs a non-zero cost to maintenance, reliability, cognitive load, and management. For example, if you are in AWS, every unused box adds time to the API call for listing the instances. Every extra box that is there for "just in case" does too. It's important to balance these things.
ryandrake 11y ago
Don't forget safety-critical software. I don't want avionics software developers thinking "that bug doesn't matter because it's very rare". Even if the effect is non-catastrophic. For certain classes of software, zero bugs is not an unreasonable or unrealistic goal.
daurnimator 11y ago
> Resources are a matter of budgetary concerns, mostly. After making my own logging system, which was expensive in dev time + maintenance, I switched our company over to papertrail ( https://papertrailapp.com https://papertrailapp.com ) they handle quite large volumes of logs for less than $100 a month. They provide an interface to search the logs with low latency. I'm not sure what type of organisation can't afford it.... (cost scales with usage)

by xorcist 11y ago

jedberg 11y ago

> Resources are a matter of budgetary concerns, mostly. True, but most people unfortunately don't have unlimited budgets to work with. Also, there is more than a monetary cost. Every system in your architecture incurs a non-zero cost to maintenance, reliability, cognitive load, and management. For example, if you are in AWS, every unused box adds time to the API call for listing the instances. Every extra box that is there for "just in case" does too. It's important to balance these things.

ryandrake 11y ago

Don't forget safety-critical software. I don't want avionics software developers thinking "that bug doesn't matter because it's very rare". Even if the effect is non-catastrophic. For certain classes of software, zero bugs is not an unreasonable or unrealistic goal.

daurnimator 11y ago

> Resources are a matter of budgetary concerns, mostly. After making my own logging system, which was expensive in dev time + maintenance, I switched our company over to papertrail ( https://papertrailapp.com https://papertrailapp.com ) they handle quite large volumes of logs for less than $100 a month. They provide an interface to search the logs with low latency. I'm not sure what type of organisation can't afford it.... (cost scales with usage)