I have sets of probing data from an internal monitoring tool which represent the availability of different services (databases, webservices and so on).
Now my task is to visualize this data and I reached a point where I should make a decision concerning the data interpretation for the most recent data set.
Quick overall definition, each data set reflects the given status at the time when it was verified and the probing intervals vary from service to service, so there is no constant time window. Hence if there is an interval of e.g. five minutes between set A and B, for these five minutes (assuming with B there is a change) the status of A is always assumed.
So far no problem, my question now however is what should be assumed for the time between now and the most recent data.
I see two solutions
- Only visualize data up until the most recent check, and ignore the passed time since then
or
- Assume the status from the most recent check is still valid up until now (like between two given data entries)
I could see arguments for both and hope someone can point me in the right direction as to which approach would be the most logical one.
1
The counter-question is, what is the apparent probing interval you want to present to your users?
If the apparent probing interval should be the actual probing interval, you should only present actual data points to the user without interpolation or extrapolation.
If the apparent probing interval should be a real-time status, you should extrapolate the values from the last probe in the same way as you interpolate values that fall in between two probes. So, if you assume a value remains unchanged until an actual probe tells you otherwise, this should also be the method in which you extrapolate the value after the last actual probe.
1