3 ms·

Ideally, you want someone who can be both while also knowing when one is necessary over the other. Is stuff actively on fire? You want a Cowboy, and you want o

Ideally, you want someone who can be both while also knowing when one is necessary over the other. Is stuff actively on fire? You want a Cowboy, and you want one NOW. You need that fire put out, you need stuff back online and work resumed. Once the fire is out, you need a Drone, someone to dive into the systems to determine what caused the issue in the first place, and ensure systems are designed to never let it happen again. For larger orgs, you want Drones on the day-to-day. Focus on iterative, incremental adjustments and improvements. A new feature here, a pipeline upgrade there, a bugfix every sprint. But you also need those Drones to become Cowboys when systems are hindering their productivity, to get something over the finish line when a system is arbitrarily blocking the path forward without valid reason. You then need those Cowboys to revert back to Drones to engage with malicious compliance on bad systems, to surface those problems up to leadership and get them dealt with. The best performers are neither drones nor cowboys, but both as needed. It's always about balance.

vanviegen 2y ago
> ensure systems are designed to never let it happen again Though that may be the right response in many scenarios, sometimes it's better to just accept the risk of it happening again, as the preventative measure may be worse than the disease. Especially when considering the combined, stifling effect of a historically accumulated wealth of such measures.
- stego-tech 2y ago
  > sometimes it's better to just accept the risk of it happening again, as the preventative measure may be worse than the disease I mean, while this is an important consideration to have, I fear that all too often it's used as an excuse to ignore the issue altogether until a later date, especially by management. Prevention doesn't necessarily mean the issue is prevented from recurring, but that the outcome is prevented. It could be building a self-healing check that intervenes on a given symptom to auto-repair a known fault. It could be firing off a warning that this issue will happen again soon, so whoever is on call isn't caught by surprise. It could be shunting a known outage into a specific maintenance window, so customers don't notice it. Prevention is ideal, sure, but sometimes the best you can do is a preemptive response.

vanviegen 2y ago

> ensure systems are designed to never let it happen again Though that may be the right response in many scenarios, sometimes it's better to just accept the risk of it happening again, as the preventative measure may be worse than the disease. Especially when considering the combined, stifling effect of a historically accumulated wealth of such measures.

stego-tech 2y ago
> sometimes it's better to just accept the risk of it happening again, as the preventative measure may be worse than the disease I mean, while this is an important consideration to have, I fear that all too often it's used as an excuse to ignore the issue altogether until a later date, especially by management. Prevention doesn't necessarily mean the issue is prevented from recurring, but that the outcome is prevented. It could be building a self-healing check that intervenes on a given symptom to auto-repair a known fault. It could be firing off a warning that this issue will happen again soon, so whoever is on call isn't caught by surprise. It could be shunting a known outage into a specific maintenance window, so customers don't notice it. Prevention is ideal, sure, but sometimes the best you can do is a preemptive response.