Server Density: Ops are not machines. Introducing HumanOps

guest post by server density,

Sysadmins are not superhuman. They get tired and stressed like everyone else.

We all know that constant stress is injurious to health. We also know that fatigue impairs our basic problem-solving abilities. So the question is: How do you keep stress and fatigue from undermining sysadmins’ ability to do their best work. over the years, and more engineers joined server density Team, we’re fine-tuning how we handle production incidents, how we handle on-call schedules, and how we work together remotely. All those efforts are focused on humans and systems alike.

In Server Density we have seen a strong correlation between human and system metrics. Less stress leads to fewer errors and increases. Fewer incidents and warnings lead to better sleep and less stress. Better sleep leads to better time-to-resolution metrics. How many interruptions and wakes do our engineers experience per month? How many late shift and weekend calls? What systems do we have in place to track this information?

With this data in hand, we can begin to consider making things better for humans. How do we reduce blockages? How do we make sure our engineers cover all the downtime and renovations they need? How can we reduce stress and fatigue? How do we create systems and devices that are intuitive, and in line with how the human brain works?

Introducing HumanOps

human operation There is a collection of principles that address those questions. It shifts our attention away from the system and towards humans. HumanOps begins with a basic belief, namely that technology affects the well-being of humans just as humans affect the reliable operation of technology.

As software makers, we have significant opportunities here.

How do you spot issues before they cause downtime? How do you reduce incidents and reduce stress? How do you present this data in a more intuitive way?

We recently launched a HumanOps feature for the iPhone called sparkline, Sparklines condense our server density charts into short inline expressions that show trends. Sparklines are the perfect match for the iPhone because they are “what’s up?” provide visual cues about. Allows sysadmins to quickly decide whether to go home, or whether they can finish dinner before reaching for their laptops.

Here is a wireframe for an upcoming server density facility is called cautious history, Notice the cost column? It measures the cost of events in actual man hours. HumanOps features like this (we have many more in the pipeline) are designed to build bridges between systems and humans. They also present information in a way that is easy for humans to pick up on at a glance.

person on call

Another important part of HumanOps is the on-call task. The anxiety associated with having hours available stems from a perceived lack of control. It doesn’t matter if the phone vibrates or not. Being on-call and not making calls is, in fact, more stressful than a “busy” shift. It is non-stop vigilance, checking for potential “threats” that is unhealthy.
HumanOps - Alert History
How do you restore a sense of control? How do you measure and track out-of-hours incidents and the human cost of escalation? All of those ideas fall entirely under the HumanOps agenda.

About server density

server density DevOps helps teams monitor their infrastructure, reduce downtime, and improve performance, all from a single console, Android, iPhone, and Apple TV app. We process over 35TB/month of surveillance data from Drupal, Greenpeace, and NHS Ambulance Emergency Response, and thousands of other organizations. Quick alerts let sysadmins know something is wrong, while state-of-the-art dashboards help them detect and troubleshoot problems before they become an incident. We offer Server monitoring and Website similar monitoring.

Server Density is available on several flavors of Linux and Windows Server. Our web interface is built using the same APIs available to everyone else. This means you get unbridled access to use all your server density time-series data however you like. start monitoring Build your entire stack in minutes with 50+ integrations, or build your own custom stack using plugins and webhooks.

Leave a Comment