Questions regarding Linode's Analytic CPU Graph
I've been monitoring my linodes' analytic CPU graphs for a while.
I occasionally noticed that there were some moments which all the graphs (CPU, I/O, Networks) went down to zero (0) for all my linode instances at the same time.
Please check the picture here
The time that arrow points to is 28.07.2022 @ 09h10 to 09h30 GMT+7
The questions are as follows…
Were my linodes stop working in that time?
Is it just a problem with system data acquisition of Linode analytic?
Were my linodes still accessible (e.g. web server) during that time?
Many thanks in advance for all answers.
1 Reply
jhartman
Linode Staff
✓ Best Answer
Although the link doesn't seem to be active anymore, this is something that I have also seen before. Essentially, we offer two types of monitoring - the free analytics displayed within each individual Linode's status page and the varying levels of our LongView Service.
Where-as LongView is installed as a client directly on your Linode, the free analytics are externally metrics gathered by the host machine. This means that in general, the LongView client is able to gather more accurate metrics (as well as internal metrics like memory and disk usage) while the analytics are just surface level metrics.
Since the analytics are gathered externally, brief networking blips or excessive host loads can briefly disrupt the graphing of an otherwise working and healthy server. Unless there is a larger-scale issue such as Emergency Maintenance for the host or a data center-wide event reflected on our Status Page, these brief interrupts do not necessary correlate with the inability to connect to your Linode.