1. For load statistics it would be nice to be able to see stats for each individual instance as well as aggregated. Say you notice that your app Role is seeing high CPU activity, if you have 5 app servers, you potentially need to log onto all 5 and run top to discover that some process got stuck in an infinite loop on one instance. (Yeah, silly and contrived example, but you get the idea.)
2. Better "getting started" documentation. I think that in general Scalr is pretty easy to use, once you understand the concepts that are at it's core. For instance some docs that clearly explain the full life-cycle of a Scalr instance, along with all of the events/callbacks that get executed, and how to hook your own processes/scripts into the ecosystem. There's a lot of "magic" built into Scalr, and that can be intimidating if you don't understand exactly what's going on under the hood and you need to tweak it a little bit. Just some pointers to and explanation of the "magic" would do wonders. Along these lines, maybe someone can tell me. What exactly are the differences and pros/cons between placing a script in the Scalr.net interface and setting it to execute on a Role VS. customizing the AMI for the role and placing scripts in the SNMP traps area? Thanks, Jeremy On Aug 21, 11:51 am, Kevin Baker <[email protected]> wrote: > On Fri, Aug 21, 2009 at 9:45 AM, Kevin Baker <[email protected]> wrote: > > > 1. Auto refresh data on logs > > 2. I like the look of your scalr.net main page... maybe use that for > > the look and feel of the scalr site > > 3. Create dashboard for home page by default. It should have widgets > > with: > > 1. uptime/status on each farm/application > > 2. general usage information stuff like: > > https://scalr.net/farm_usage_stats.php > > 4. Optionally allow us to set main page after login to goto any of > > the pages > > 5. For new users give them more instruction on setting up a farm > > 6. Wizard for setting up DNS early on. The DNS was essential for us, > > some recommendations would be helpful. > > 1. Setting up secondary DNS > > 2. Using a dot info domain with scalr and then using cnames to it on > > our primaries.... much eisier to manage. > > 7. more to come.... > > Much more robust monitoring: > - on call rotation > - roles/groups to alert specific roles based on specific events rather than > specific email addresses > - SMS alerts... we set this up with a third part gateway for cheap, be happy > to give details on this > - Alerts on failed http requests, that work the same way the e-load balancer > does. > > Scalr on anything: > - it would be great to switch the way scaling works a bit. I really like the > voting system that rightscale has. Basically any event can vote to have a > server scale, if there are enough votes, based on configurable threshold, > servers scale > - be able to write custom scripts that can submit votes on anything. > - if not the main thing we are hoping for is to be able to scale based on > the "http response time" of a service. > > > > > 1. > > > 2009/8/21 Frédéric Sidler <[email protected]> > > > Here are mine ;-) > > >> 1. Look at the statistic usage. Not only the farm statistic usage, but > >> also the role statistic usage. Unfortunately the the page to the stats > >> of a > >> role is not permalinkable ;-( This is a pain for me. > >> 2. Don't forget that from time to time we need to deploy new version > >> of the application to our servers. Some changes can be done without > >> stopping > >> the service (app role update), but some changes require to put a > >> temporary > >> "down for maintenance" page. This could be done via nginx I think with a > >> simple scripting procedure from your part. > >> 3. You were asking for some premium services people would be ready to > >> subscribe. Here is one. My instances are managed by Nagios. Nagios is > >> great, > >> but it is a pain in the ass to manage instances with Nagios especially > >> when > >> new instances are added automatically to the farm. You have a great > >> email > >> notification service, but your monitoring solution only take the > >> instance > >> into account, not the service and different parameters that interests > >> me. > >> You know, I know that Rightscale uses syslog-ng. I would like to have a > >> monitoring service that is able to monitor > >> 1. disk space > >> 2. personnal processes (Jabber, SMS, Postfix, Mecached) > >> 3. we need to be informed of what is going wrong on our servers, > >> before our users even notify us of the problem ;-) > >> 4. Now I'm used to, but it was confusing to call the load balancing > >> role "www" at the beginning. Because "www" was associated with the web > >> server for me. This is not for me, but for the new users. > > >> On Fri, Aug 21, 2009 at 8:48 AM, Sebastian Stadil > >> <[email protected]>wrote: > > >>> Hello everyone, I think we can all agree that Scalr could be made easier > >>> to use, so we're starting to work on an easier, more accessible UI design. > > >>> - The Wizard guides you from start to finish for creating an > >>> application such as a website > >>> - Covers setting up initial DNS, > >>> - Setting up Resources/Farms, Instances, and Roles, > >>> - Setting up your Application, option to pull from repository, > >>> - Application versioning > >>> - The Top Menu lets you perform individual tasks, among which are > >>> - Maintenance of infrastructure, such as monitoring and taking > >>> action* > >>> - Maintenance of your applications, such as application versions, > >>> and the possibility off rollback to a previous state with whole farm > >>> snapshots > >>> - Adding Farms, Applications, Subdomains, Whole domains, Scripts** > > >>> * What sort of actions do you usually take after looking at the > >>> monitoring page? > >>> ** What are the individual tasks that you use the most? Which ones aren't > >>> listed / would you like to have added? > > >>> We'd like input from the community. Please respond with a task that you > >>> often perform, and a suggestion on how to streamline it. > > >>> Cheers, > >>> Sebastian > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "scalr-discuss" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/scalr-discuss?hl=en -~----------~----~----~----~------~----~------~--~---
