I would like a current usage monitor, and bandwidth and cpu and ram chart.
I would love to get the droplet graphs as a part of the API. I've looked through the API docs but I cannot find anything that would let me get bandwidth, cpu, and disk usage.
It would be great if there is feature like Bandwidth limiter, it's mean if bandwidth usage reach XX% so the interface gonna trigger some command like shutdown or something,...
Bandwidth should calculated on aggregate basis rather than droplet base. otherwise the unused bandwidth in private networking will be injustice.
Please, anyone interested in make this even better, vote up here: https://digitalocean.uservoice.com/forums/136585-digitalocean/suggestions/6092926-better-droplets-dashboard :)
It would be awesome to know how much / what % of total bandwidth I've used vs what is remaining at any given point during the billing cycle!
+1 for adding RAM usage as well.
Another user who would love to have chart that shows total used bandwidth or the amount of data transfer remaining....not just the megabits being transferred at a specific point in time. :-)
This is a great feature -- would be nice to get RAM usage as well
" # free -m " - showing on 70 MB of my 1GB RAM is free, though my droplet is just sitting idle. In the graph I can see 0 traffic , almost no cpu and disk usage. I wish I could have attached the screen shots here :)
The graphics do not make clear the data.
I suggest to improve the identification of graphs.
Any progress on memory usage?
RAM graph would be excellent.
(Active; Buffered; Free)
What about the RAM usage? Still unavailable?
The graphs are based on a 100% scale so as you approach that that means you are running out of resources. The CPU is reporting system and user usage so user = applications that you are running like php, mysql, etc.
As for bandwidth graphs we will most likely not be providing that until we roll out private networking in each of our datacenter regions. For the time being while we aren't providing bandwidth stats we also aren't billing for bandwidth.
ETA for total Bandwidth usage in the month?
And with the graphs it would be best if a best standards or best performance is given. so that I may understand the graphs and know what to modify if needed.
You can use New Relic: http://newrelic.com
I should also mention my server seemed to be down during that period; The site was down and I could not connect via ssh so ended up having to click "Power Cycle" to get it back up. The user CPU appears to have returned to sensible levels after the reboot.
Great stuff! Although I should mention I had a long period today where my "user" cpu usage appeared to be up 80% between 16:30 yesterday and 10:00 today. What exactly is the "User" graph indicating, and any idea why it may have been so high for so long?
Would be great that the bandwidth usage be included in the API =)
Nice! Now DO provides network stats and droplet resources usage over time. Probably one of the most wanted features! Good job, and thanks for the updates Mr. Uretsky.
It is storing data up to a year.
And we'll be refining the underlying architecture over time.
Bandwidth usage in terms of billing will be in a separate display from the network graph which shows live utilization in mbps, and usage is in GB.
Is the system able to store all our graph datas or it will just shows certain period?
Total Bandwidth usage? I only see Network, Disk and CPU. Thanks for the update.
Deploying now =]
RAM is unavailable at the moment but we will be adding it later, we want to see what the load looks like as it's rolled out to all users and collect feedback. =]
Still not available for new accounts?
Looks cool, but does not show RAM usage. Is this possible to add this?
It takes a couple seconds to load the graphs (Latest Chrome Stable), you could show a loading indicator while it's loading.
@moisey it's working very well for me. :)
This feature has been rolled out for grandfathered accounts and will appear under a new icon in the droplets toolbar called "graphs"
Looking for feedback, thanks guys & gals =]
Any recent updates? I think this is the most wanted feature right now :)
Thanks for the new update Uretsky :D
I'll provide some basic details as things are still changing during the roll-out.
We are collecting stats from the hypervisor layer so there are no agents or collectors that need to be installed or running on your droplets.
We are collecting: network transfer, disk I/O, and CPU.
We are running our own aggregator that then compacts the data into various intervals.
We are working on providing:
4 hour graphs (1 minute interval)
1 day graph (5 minute interval)
1 week graph (30 minute interval)
1 month graph (2 hour interval)
1 year graph (1 day interval)
We are using the D3 graphing library for display. It takes quite a bit more work to integrate but will provide the greatest amount of flexibility moving forward.
The first graph that will be displayed is network, followed by disk, followed by CPU.
We are pushing hard this week to get the first graph, network, deployed this week for grandfathered accounts as a limited beta test to see what the load will be like on all of the stat collection and display servers.
Moisey Uretsky, could you please share some details of what you are using to get the data, to store and aggregate it? Most of us like to know about nosql and bigdata.
Highcharts coming soon =)
Thanks for the update =)
P.D: It also would be interesting adding this feature on the API...
Thanks for the update. :)
We are going to be rolling out the first iteration of our stats service to our grandfathered customers!
We are going to start with bandwidth graphs and we are also limiting the initial roll out to collect feedback, bug reports, and also to stress test the underlying system to see how it will handle thousands of requests so that we can tune it.
We'll keep everyone updated this coming week as this feature starts to go live and as we collect more information.
Bumping for updates :)
any updates ?
Can't wait to see this feature working! Specially bandwith graphs.
We've finished the first phase of stats collection which is now running in each region.
We are building a service which will export that information out from which we will be building the graphs so I think the first public beta of this will be appearing in the next 3-4 weeks.
Looking forward to this!
The larger component is the stats collection which is nearing completion, after that we'll be able to use a graphing library most likely D3 for presentation.
So for now there is nothing for end users then. I got all excited and went to check before I had even read your last comment completely lol. I assumed since my VPS is in NY that I would see stats somewhere. lol.
Wanted to provide an update on this item.
We've rolled out stats collection to all servers in our NY datacenter and are playing around with the data store and observing how it handles the load and frequency of data.
Once we've had a chance to sort out the bugs and fine tune things we will be rolling it out stats collection to our other DC and then we'll be able to push that out into graphs for the front-end.
Eventually we will also provide this information via the API for polling directly as well but that may be delayed until we roll out APIv2.0.
yes I want too.
Yay! Also available in the API
would also love if this is somehow implemented in the API
+3 for this, also good to be able to access that through the API
What about Gangliat
Yes, very good. API to these stats is very important to me as well.
IMO, this is the only major thing that's missing. You guys even have remote vterms (which is pretty awesome), but usage stats are absolutely essential.
I think DigitalOcean differential is to be easy of use. This feature will save a lot of time.
+1 it would save us from installing munin / zabbix.
You can install munin package and watch all theese stats
Mail me, if you ned help
Statistics and things like bandwith would be nice.
Much like everything else it's about iterating and refining. =]
So you bill by the hour up to the month cap. How will bandwidth transfer limits work? If I use on a $5 server all 1TB in a week, and then create a snapshot of that server, destroy it, and launch a new droplet for the rest of the month and don't go over the 1TB, will I only be charged $5? How are you going to make this work and at the same time not get abused?
Until the bandwidth usage is displayed we wont be charging you =]
Win - Win
I am a big AWS user, and love digitalocean.com... but if DO is going to charge me for bandwidth usage, then they need a screen where I can see how much I've been using.
FWIW, we use New Relic's server monitoring to show all of those stats and it works great. It's free as well.
Just an FYI, linode uses munin to output graphs about that usage, which is pretty cool. I like graphs.
This will be an extremely useful feature, most VPS companies have that already but I would rather use this company and not have the resource monitor. Since they are working on it this is a bonus!
This is on the horizon and we'll be starting work on it soon.
or using *nix command like that : df, df -h, df -i, free -m, ps aux
We're working on stats integration into the front-end, right now its taking a backseat currently to some heavy lifting that we are doing on the backend and physical infrastructure but that should be finished by early January.
Then we'll back on track with front-end updates and this will move up the queue quickly.
I have setup my first VPS and installed a WHM/cPanel as my control panel. If you folks would like a cheap WHM/cPanel license let me know and I will provide with the link.
It's really awesome after setting up my VPS using WHM/cPanel.
I would love to know how much space im using all together on the server.
On a *nix system you can use webmin with webminstats to achive that.
The graphs under the "Statistics" link are nifty-looking and all, but it might not be a bad idea to pull them from the production environment while they're just displaying static demo statistics that are completely irrelevant to users' droplets. Just put the demo on a "features" page and mark it as coming soon, so people can see what it looks like if they're so inclined, but understand that the info isn't available yet :)
We're working on graphing but we're hoping to avoid doing the standard MRTG, instead we want to make something a bit more interesting that will hopefully allow users to get some performance analytics without needing to resort to something as intricate as NewRelic.
It's in the pipeline but we're currently focused on a few larger features. This is definitely on the roadmap though.
PS> I think I know which competitor you're referring to ;)
You won't be notified about changes to this idea.