Manager is Latin for…
Not many people know this but the word ‘manager’ comes from the Latin verb meaning ‘to monitor your people’.
That opening line is, of course, complete crap. But it does hint at an important and often overlooked part of what an SRE or Operations manager’s job entails.
Businesses will instill that the main tenant an SRE team should follow is this: Keeping production up and running is the most important thing above all else.
After all we build in scalability, resilience, high availability and self-healing as part of the platform. All with the intent to minimize any potential downtime. We invest time and effort in getting observations and monitoring correct, to ensure that nothing is ever missed, that potential problems are caught ahead of time and addressed accordingly. The smallest of alerts fires to give the SRE team a heads up that something bigger could be brewing. But there is one very important component that no amount of tooling or redundancy can be used for: monitoring the team that looks after production.
Now, let’s clear up a point here before I go any further. I am not saying that people and components are interchangeable objects in the world of IT. Components are the hardware that your platform runs on, the virtual machine that hosts the software. People are none of these things, but they are a component that needs to be monitored when viewing the overall health of your platform nonetheless. After all, production environments don’t just appear overnight. They require a team of people to bring them into reality. If that team disappeared as soon as the platform went live your company would be in a very dangerous position.
Which is why a manager must monitor the people in their team. Not monitor in the sense of ‘are you doing work there you lazy slacker’ — rather monitor the members “health” just as the monitoring tools monitor the ‘health’ of the platform. Now, to poorly paraphrase Bones McCoy, “For God’s sake, I’m a manager, not a doctor.” — meaning I am not advocating that managers suddenly start doing physicals and taking blood pressure. But you need to remember that a component in production has a health check to indicate it is working well and not suddenly going to fall over, similar “health checks” can be applied to your team.
What do these ‘health checks’ typically look like? Well they sound very similar to the following questions: Are they under too much pressure? Have they been a little short with you or others, which is wildly out of character? Are mistakes being made on fairly simple things? Are they complaining more these days than the normal ‘Why do we have to work?’ levels of complaints?
These are all your ‘alerts’ going off that something isn’t right with this particular ‘component’ in production. The problem here is that there is not an entire team to investigate the issue and resolve the ‘alert’ — you’re it. The manager. You need to act in order to resolve that ‘alert’ before it becomes a bigger ‘incident’ that you cannot deal with.
Luckily the troubleshooting steps are fairly similar. Acknowledge the alert, see what may be causing it, take corrective steps to resolve it and stop the alert happening again.
Maybe somebody is feeling pressure because multiple projects require their involvement for very short timelines. Possibly something is going on in their personal life and it is spilling over into their professional. Your corrective action here? Have a chat, see what’s going on. You probably cannot help resolve the personal stuff, but you can divert work away so that maybe they have a bit more head-space to deal with things. Likewise if they are feeling under pressure, a simple chat can help them to understand which tasks are priority from a management (your) point of view and takes a bit of pressure off their shoulders. Or you can escalate up to Upper Management and say that the timelines are too tight and causing pressure so something is going to slip by design rather than accident.
That last step is actually a crucial one to not be afraid of. Timelines for projects can sometimes be agreed in complete isolation to other work being done in the company. One Manager-type unaware that another Manager-type has committed to something at the same time will always result in Chaos Management when it comes to resourcing for projects. People cannot multi-task, it is a fact of life. The problem is the further you get away from monitoring and being in the operations battlefield, the easier it is to forget that in order to avoid large incidents you have to react to small alerts fast.
Ways to avoid causing these alerts from firing involve all the things that a person will say ‘Well that is obvious!’. Such as no last minute change windows, repeated late nights or unrealistic timelines decided in complete isolation. But if, for reasons beyond your control, you cannot do this then when such ‘alerts’ fire you must bring them up the command chain. Otherwise that ‘alert’ is going to keep firing and you’re never going to get it resolved.
Which, let’s face it, is just never going to happen when you’re an SRE at heart because no alert can be left unresolved. Anything that fires will have you looking at it. All you can do as a manager is try and stop the alerts from getting to a point which causes an incident.
But what is an incident when you’re talking about ‘monitoring’ people. Simply put it is an exit from the company. If enough alerts fire in your pool of talent they will, eventually, take corrective action themselves. Since their expectation would have been that management fixed things for them, your resources will take the only step that is utterly in their control: to no longer work in a company that seems to value 14 hour days over work-life balance.
So generate your reports and do your quarterly reviews, these are all ways that can alert you that something might be wrong. But having a non-work related chat every few days with people is a surefire way to let a good manager monitor their team and see that the health of the most important component in production is green across the board.
Hilariously the probable origin of the word ‘manager’ comes from the Latin ‘manus’ meaning ‘hand’. A good manager provides the necessary “hand” for their employees when they need it, to ensure those ‘alerts’ don’t fire.
