New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
You should request they assist you setting up your infrastructure of highly available failover idle balancers.
The thing about Automatic Vertical Scaling is that each of your containers can scale down ~128MB RAM ( 1 cloudlet ) automatically when not in use and you pay almost nothing for them. You can also set it horizontal scaling to remove nodes when load is less.
Best part is no fees for shutdown / stopped environments . As long as you are within the free 20GB limits , have no public IPs or licenses, you can indefinitely shutdown an environment at no cost. A number of developers use this for Dev environments . We even have automatic start / stop scheduling where environment can wake up at say 9AM and shutdown at 8PM when you finish work.
oh, I am sure there is a lot one can do with this - did not want to argue that ;-)
I simply haven't discovered my real use case for any kind of that cloud-scaling stuff yet. esp. without just burning money on it
Will definitely try this
Signed up since bandwidth is fair now. Thanks for the quick and better move. I can use your platform to scale up my app server, when more users are engaged at the same time.
I've upgraded my account and got $10 credit, can I still schedule a demo (and possibly get $40 extra credit)?
that's awesome!
Yes, you can.
Ok, so I just did a demo with @leapswitch and cloudjiffy was better than I expected. It was more of a managed VM rather than just a bitnami pre-installed VPS you can get from the normal 'clouds'. You can edit any configs, etc. like you'll normally do on a VPS while the system handles all the sysadmin jobs like HA/Replication/Synchronization/etc.
The only downside I can see is the reserved cloudlets being a bit more expensive than what you will normally get from others, and lack of locations. Which I do believe can get better when the scale gets up.
I'd say the most important part is that the guys behind this actually listens to customers and makes changes accordingly (like the bandwidth issues we had on this thread), and they are really friendly and helpful.
Overall, I'm very impressed of cloudjiffy, and if you haven't done the demo, I highly recommend you do, and also take advantage of the $50 (idling) credits.
Edit : I forgot the most important part, money. The billing may be a bit different than what you'd see from a regular provider, but if you take a closer look, it's pretty straightforward and you can save a lot if you take advantage of it (and heck it's way more simple than AWS/GCP/etc.). Since most of my stuff idles, i did a test run and most of my stuffs spiked a bit at initial installation, and after that idled at the minimum 'reserved cloudlets' (which is 1 cloudlet ofcourse).
kudos to Ishan from @leapswitch for taking the time and doing that demo stuff!
seems like you built a great product/platform over there. gonna give more feedback after playing around with it and burning the credits ;-)
keep up the good work and your friendly, supportive attitude!
Damn, why no gameservers? I've just seen that and I already deposited $10. I was hoping to test it out, and game servers are the best organic test there is! Variable player count, resource usage, RAM needed and stuff. Well, too bad.
Do you have a refund policy? I looked through your TOS & thread but found no mention of it.
I guess it's due to the potential DDoS attacks that a gameserver might attract. However I do agree gameservers can very well take advantage of Cloudjiffy. Maybe a second thought @leapswitch ?
pay 10$ and I get 30$ ,thanks!
I have not found this feature in settings, how to do that?
how much it costs if one cloudlet runing less than one hour?
You are welcome
You can find it in Marketplace > Addons or in any environment > Addons.
https://cloudjiffy.com/portal/knowledgebase/212/Environment-StartorStop-Scheduler-Add-On.html?language=english
We take usage every 20 seconds and average it over an hour. So if your average is > 0.5 cloudlet it will be rounded off to 1 cloudlet. Below this it won't be charged.
They are not allowed due to DDoS attacks. I have reached out to you via PM to discuss your use case further.
We refund any unused paid balance .
My balance is showing up as 50 INR instead of USD!
Please use https://app.cloudjiffy.co to access ( .com is for Indian/INR customers)
geekbench for an elastic VPS in FRA, set to scale from 1 cloudlet up to 12 (max 4.8GHz, 1.5 GB): https://browser.geekbench.com/v4/cpu/14298807
the statistics are helpful to estimate, what the environment really consumes:
the last bigger spike in CPU/RAM is when the geekbench ran...
also ofc download bench for FRA - network is fast my man ;-)
on a second thought, I'd say that's beyond the call of duty ;-)
if possible and to protect against misuse you should look into still calculating the free traffic based on the runtime of the environment. maybe you can somehow calculate any overusage not only per month but for used hours on the point of deletion of an environment etc.
so if I remove an environment after a 40 hours I am fine with e.g. 40G free traffic for that instead.
Thank you for the report and feedback.
This will take time to implement. We will be implementing the free 730GB option first and then look at implementing it on a per hour accumulated basis.
I found out that 1 x 100 (max) cloudlet is worse than 5 x 20 (max) cloudlet.
When using Max 100 cloudlet my apps is not responding when used by a lot of users. Although the max is still not reached (I am talking about 10GHZ++)
It's much more responsive when using multiple web server. Any one can explain why is this the case?
It could be the software you're using is not utilizing the resources properly. It could also be a network/disk io problem
I am using their template (apache) and WordPress. I don't think io is the case because they both using the same storage cloudlet.
Maybe I am not clear enough, 1 cloudlet with 100 vertical scaling, perform much much worse than 5 horizontal scaling and 20 vertical scaling cloudlet.
Is it because one web server is considered 1 core. So when I use it, it's considered as 1 core server with very big clock (E. G. 20 Ghz)?
It seems like you get totalGHz/coreGHz numbers of cores (e.g. 11GHz gives you 5 of 2.2 GHz cores). Not sure why vertical scaling performs worse than horizontal though. Afaik in theory vertical should work better since no clustering overhead is required (although it seems like you used the clustered WP template, which should make vertical and horizontal perform about the same)?
You can run
cat /proc/cpuinfo
On web SSH to see core counts, speeds, etc.
In Apache we have seen MaxRequestWorkers being hit ( even though we auto tune it based on resources). You can manually edit MaxRequestWorkers ( remove auto configuration mark at the top of the config file) and try it again.
No, I used Their PHP Template (Varnish, Apache, MySQL, and Storage), and install Wordpress manually.
I though so too. That's why I set the vertical scaling limit very high. Until someday my apps failed to load. And then I set the horizontal scaling, and my apps become much more responsive. The only bad thing about horizontal scaling is they took time to scale, normally about 30 sec. The worst case is when the load is very high, sometimes it failed to scale.
Tough, The high usage and horizontal email notification is neat. You know to keep monitor your service when there is notification come to your email.
Noted, I might try this in the future in busy month. As of now, mostly my server are idling, from 7AM - 10PM. Gotta save some pennies (9 hour offline / day = saving 270 hour /month).
How many cloudlets is MySQL using when on idle?
I believe it's 4.
Are you using replication of any sort? Mine's going between 3 and 4 after a spike (does goes down to 2 after a reboot), with no users, 1 plugin, 1 post