Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials

Enhancement: adjust scale on graph per monitor/resource per view

I don't know how APM determines what scale to use, but most of our graphs appear as if the answer is 0 because the value is in milliseconds and yet the Y axis has a high value of several million. It would be helpful if we could adjust the Y axis max so we can see something useful.

Find more posts tagged with

Accepted answers

All comments

VáclavSáblík

Hello,
Could you please specify more exactly what kind of chart you are mentioning. I was looking over several of them and for me where top Y margins giving sense. For instance response time that is 0ms all the time has maximum 1000s which is not enormously big number.

I will make sure adjusting Y axis limits will be logged as a feature request but we would like also to keep automatic adjusting as much useful as it could be.

Thanks,
Vaclav Sablik
SolarWinds Developer

MrSampsonite

HI I was talking more for any graph. For example we are monitoring several perfmon counters on our servers. For some there are spikes so high that it makes reading the normal values impossible. If we could adjust the Y axis to be just above where the average utilization is we could probably make out what they are visually (instead of exporting and running Excel functions).

If you're in the mood for submitting feature requests I have a couple to give. I try not to bloat the feature request forum.

One is the ability to select multiple monitors at once in a template and delete them or copy them. It takes forever to delete one, let it refresh, etc. And then at the end we have to click submit. If the page timesout for whatever reason we just lost all the deletes. And often it's much simpler to copy a monitor multiple times and just change a few things in it. I love that feature in "customize this page".

Secondly, back to the graphs, it would be so helpful to do the following two things to graphs:

1. For each shown graph, in the text below it show the average of the statistic. The 95th % is helpful, but for many just setting the time period and seeing the average saves a TON of time compared to clicking the graph, exporting to excel (if the PC has excel or openoffice on it) and then highlighting rows and running functions on it.

2. Secondly, I don't know about other people but we have a bunch of graphs in different views. For example we have a "disk stats" view so we can see perfmon stats for various disk counters. What would be fantastic is the ability have a master drop down which lists various (or custom) time frames and then once we select it, it either adjusts all the graphs to the same time frame or we can select which of the graphs on the page to make this adjustment too. Right now I have to go to "EDIT" on each graph and change the timeframe of each before I can get a consistent view. To change from last 4 hours to last 24 hours or last 7 days or last "custom" time wastes valuable time.

Anyway thanks for asking. Hopefully I didn't go overboard.

VáclavSáblík

You are right that for spikes manual scale is good approach.

Deleting multiple components and setting time frame for whole view or subset of charts is something we are definitely considering.

Speaking about average, I was under impression most important are spikes i.e. when something goes different than expected. Information you are looking for average is sort of new to me. Especially when it is so important you are forced to copy data to excel for average calculation.

I’ll pass this through to PM; your feedback is very valuable.

MrSampsonite

Spikes are important if they are over a significant period of time, but with large SQL operations (the types of servers we're monitoring) spikes are normal if they don't last long. Of course the time period matters. If I'm look at last 4 hours and a spike takes up 1/4 of the chart then it's not a "spike" per se. But something to investigate. But if I look at last 1 hour and I see a couple spikes here and there that last almost no time, the average is much more reliable. Anyway, I imagine calculating the 95th percentile is more resource consuming than AVG(stat). I don't know if we're an abnormal customer but AVG is the typical value we're basing conclusions on. There are charts that show min/max/ave but that's on a per data point basis. Seeing that value over the period of chart would save us tons of time. Otherwise I'm using excel or creating new reports for every time period I'm asked about.

Thanks!

epenney

Same issue with charts - 1 single spike - which in our case is due to Daylight Savings time, which results in 2hrs worth of data collection compressed into a single hour. Results in long range charts, ie: > 3months, alters scale and affects readability of chart. It would be really nice if NPM offered ability to "Ignore Overflow Value" or something to that affect. I believe PRTG offers this under advanced features.