- or
No existing idea results
- ~ No ideas found ~
184 results found
-
Disable asset pre compilation on servers not serving web content
We have separate process server which is not used to serve any web pages at all. However, I've noticed that during deployment it performs assets pre compilation which slows down deployment and is not necessary for this server. I'd like to be able to disable asset compilation through server settings.
Something likecx server-set MY_STACK asset_pipeline_precompile false
6 votes -
Temporarily disable some rake tasks
It would be convenient if a scheduled rake task could have an "enabled" flag.
In the event of a primary stack failure, a failover stack could have all the rake tasks already entered (but disabled), and it would just be a matter of disabling the primary stack's tasks and enabling the failover stack's tasks. Right now, you have to manually add and delete them.
3 votes -
8 votes
-
Make your UI more friendly
I feel lost when I access my dashboard at cloud66, is that because cloud66 is targeting experienced web masters?
1 vote -
Validate SSL Cert when enabling SSL Termination
An incorrectly formatted SSL cert, really easy to do, seems to stop ngnix from loading. Which causes a non-obvious failure. Validating the SSL cert when added would help prevent this.
1 vote -
Restart workers on stack restart
When calling cx restart on a stack, the workers aren't restarted.
This is surprising when the "cx restart --help" docs state " This will send a restart method to all stack components. This means different things for different components."
My scenario is as follows. Deployed, but needed to run migrations. Ran migrations. Ran cx restart. Web picked up migration changes, but workers did not because they weren't restarted.
Currently the only ways I've found to restart a worker is through the web interface or via ssh.
4 votes -
Bahnhof Cloud provider support
Bahnhof is known for their security measures and would be a welcome Scandinavian provider. For those who take security and privacy into the equation (next to performance and price) Bahnhof would be a great option.
3 votes -
Allow per-stack (adn per stage) keys
Allow to generate a set of deploy keys per stack and stage so these can be added as deploy keys in GitHub.
Organizations / Admins have issues with adding a single key to their user account, as a compromise of that key would result in a massive leak.
6 votes -
Support Postgres SSL or SSH tunnel connections from app servers
Support the application servers connecting to Postgres over SSL (preferred) or even over a SSH tunnel.
25 votes -
Freeze/Suspend a stack
We have a BETA stack which we bring up every now and then when we need to test new features before deploying to production.
It would be great if it was possible to freeze/suspend a stack.
I imagine hitting freeze on the stack, it takes a full backup of my database, save the current git repos id, cancel my servers (or ask me to do it). On a later time I can hit the unfreeze button, it will then setup new cloud servers, install software, use the specific git repos id, restore the backup, and setup the elastic address to point on the new servers. So the stack is back to life.
It would make it easy to safe money on servers/stacks you only use rarely.
Cloud 66 can then charge us a small amount for storing the data while the stack is frozen.
We have a BETA stack which we bring up every now and then when we need to test new features before deploying to production.
It would be great if it was possible to freeze/suspend a stack.
I imagine hitting freeze on the stack, it takes a full backup of my database, save the current git repos id, cancel my servers (or ask me to do it). On a later time I can hit the unfreeze button, it will then setup new cloud servers, install software, use the specific git repos id, restore the backup, and setup the elastic address to…
10 votes -
If there is no heartbeat on a server, it would be nice if it tried to reboot it.
If there is no heartbeat on a server, it would be nice if it tried to reboot it (using the cloud provider's API). Suggested by Jevin Maltais.
1 vote -
create a global gemrc
Create a file called /etc/gemrc with this in it:
gem: --no-document
this will skip the useless gem docs and speed up gem installation time by 300%.
1 vote -
Support for MariaDB
It should be pretty straight forward as it is a drop-in replacement for MySQL. MariaDB is often performing better than MySQL and community improvements from Facebook, Twitter, Google, and Percona all roll into MariaDB sooner than they appear in MySQL.
16 votes -
Don't automatically enable processes
When a Rails app is deployed for the first time with a Procfile (and I think also when a processing server is added), 1 process is automatically booted for each entry in the Procfile.
This is bad for at least two reasons:
Testing - processing may begin before you can verify deployment was configured correctly (e.g all correct environment variables have been set, connecting to correct DB, etc)
Resource exhaustion - deployments with many Procfile entries could lead to lots of unnecessary connections being made to databases, and preventing other applications from accessing these resources.
We have about 3 dozen Sidekiq queues with 20 Procfile entries targeting specific queues (in case we need to temporarily divert processing). Deploying a new stack will immediately spawn 500 database connections (20 process X 25 Sidekiq workers) exhausting our allocation.
It would make more sense to allows us to scale up the processes as we need them.
When a Rails app is deployed for the first time with a Procfile (and I think also when a processing server is added), 1 process is automatically booted for each entry in the Procfile.
This is bad for at least two reasons:
Testing - processing may begin before you can verify deployment was configured correctly (e.g all correct environment variables have been set, connecting to correct DB, etc)
Resource exhaustion - deployments with many Procfile entries could lead to lots of unnecessary connections being made to databases, and preventing other applications from accessing these resources.
We have about 3 dozen…
3 votes -
Support for Galera Cluster for MySQL
It would be great to have the option for using Galera Cluster to create a true Mysql cluster instead of using master/slave replication.
9 votes -
2 votes
-
Set the nginx http to https regex in the manifest file
I would be awesome to be able to set you're own rewrite regex in the manifest file. Especially in conjunction with setting redhttpto_https to true.
That would enable you to use something like ^((?!assets).)*$ to redirect everything except assets. It would make it a lot easier to have everything preconfigured for cloudfront.
1 vote -
set http to https redirect in manifest file
It would be helpful to be able to set redhttpto_https from the manifest file so a new stack could load with that setting.
1 vote -
Flowdock support
Support notifications via Flowdock Team Inbox https://www.flowdock.com
8 votes -
allow deploy hooks on haproxy instances
I'd like to automate things such as installing newrelic monitoring when spinning up haproxy load balancer instances. Currently, I do this using deploy hooks on application servers and it works great. If I could do the same on haproxy servers, I'd be thrilled.
3 votes
- Don't see your idea?