Maintaining server performance: my routine with Digital Ocean

Greeting Steemit Family

Sometimes you might be lucky and not have challenging tasks at your workplace. But as a new month starts, you have to begin all over again, while new tasks are introduced as the month unfolds. This is common when you are a service provider, and you often have to repeat the process for different clients.

At my workplace, my monthly routine involves monitoring all our running servers. This might look easy when none of the servers have problems, and you browse around, it’s over quickly. But it can be very challenging when you have to solve a bug that you are facing for the first time. Coming across familiar bugs is never a problem, they are always easy to sort out.

Monitoring servers is a whole task on its own, which is broken down into steps. I am more familiar with Digital Ocean droplets (servers), and as part of my monthly routine, I go through all our droplets to monitor performance, security, and stability. In doing this, I have to pay attention to some key areas and tools when picking a particular droplet.

I start with System Resources, which are easy to understand from a distance point of view, under the graph section:

CPU: Here, you try to figure out if the server is too busy with too many requests relative to the workload. I pay attention to the server’s CPU percentage because a very high percentage can cause the server to slow down or become unresponsive. At this point, some projects might experience downtime, which is not good for our clients.

Load Average: Here, you see the number of processes waiting for CPU within the timeframe of 1, 5, and 15 minutes. If the CPU is fully utilized for a long period, the server is overloaded. Nothing to worry about when it’s between 1 and 5 minutes. By the way, these timeframes can be identified by different colors.

Memory (RAM): You know the task of RAM—having more free RAM leads to a fast, smooth server. If free RAM is low, the server might crash at some point due to too many programs running.

Disk I/O: This is not about on and off; it’s about read and write speeds. Here, you look at how fast data is written or read from the server. This is especially important with databases, and if the speed is low, applications on the server will take a long time to respond.

Disk Usage: Just like your personal computer’s hard drive, the server stores files, logs, databases, and applications here. If the disk gets full, the server will start experiencing downtime. One of our servers had this problem, and I had to resize the droplet to increase resource capacity.

These are the basic things anyone can check to figure out if a server is healthy and running smoothly. Under no circumstances should you skip a week without checking these areas if you are a server administrator.

---

---

Cheers
Thanks for dropping by
@fombae