Junos OS devices can be located almost anywhere: in general office space, on a rack in an equipment room, or as part of a large “farm” of related devices, sometimes the size of a factory. But in all these environments, the Junos OS device typically hums merrily along with no one watching it. But what if the hum turns into the software equivalent of grinding brakes on an automobile or of a weird ka-chunk whenever you shift gears? In that case, it's very important to be able to know when the problem started and what else might have gone on during the same timeframe, along with having various other details available if needed.
In other words, you need logging. As you probably know, logging is how the system keeps track of what it's doing in a human-friendly way.
You can set up two types of logging on a Junos OS device to record events as they happen:
Although you may be tempted, you don't really want to watch and record everything with maximum syslogs and every possible tracing function (because much of what happens on a network is, thankfully, relatively boring and repetitive). Instead, you really need to look for events that indicate things may be going haywire, and then focus on just the areas that might be the cause.
For this reason, a lot of operators set up syslog ...