31 lines
1.5 KiB
Markdown
31 lines
1.5 KiB
Markdown
# Service Monitoring
|
|
|
|
Several types of monitoring:
|
|
|
|
## Building Monitoring System
|
|
|
|
If your monitoring a lot of _stuff_ know that you will could have lots of stuff going across the network(in some cases).
|
|
This is why a good monitoring system shouldn't get in the way of _business traffic_ as the purpose of the network is to service its users.
|
|
Ultimately however to ensure that things are running smoothly some monitoring must be put in place which means we must use valuable resources to make sure things are behaving correctly.
|
|
|
|
How this is done is generally up to implementation but there a few rules of thumb:
|
|
|
|
* Don't query everything on the network all at once as you could easily clog the network and drop all kinds of data in the process
|
|
|
|
If you must query _all the things_ then try to stagger the query to a few machines at a time, either by scheduling the checks or doing them randomly to avoid them all at once.
|
|
|
|
* Limit what kind of data you query for
|
|
|
|
The reasoning is the same as before, try to not use too many resources at the same time lest you risk dropping both business and query data
|
|
|
|
* Avoid peak times
|
|
|
|
All networks have certain hours of the day where usage is at its maximum for the size of that userbase. If you're dealing with a rather large network then definitely avoid doing heavy monitoring tasks during these times.
|
|
|
|
|
|
Know that we have to _get data, store data, and ultimately view data_.
|
|
|
|
## Historically
|
|
|
|
SNMP: Usually you will get back a ton of data
|