- add rancher haproxy support for hostnames with :80 at end
- add graphite exporter to server
- expose graphite port
- fix --port => -p in rancher server bootstrap
- run server with bridge network instead of host, add graphite options
- add monitoring and elasticsearch nodes to prod deployment
- remove mongo images (now at https://github.com/finboxio/mongo
- make global staging spot price configurable
- add slack channel to prod mongo infrastructure
- update for server/agent image specification, bootstrap image version, and slack channel
- fix ami variables
- propagate rancher_server, rancher_agent
- fix reference to rancher hostname in status infra
- propagate rancher_server, rancher_agent
- propagate slack vars/defaults
- make more slack options configurable (channel, username, icon)
- make rancher server/agent images configurable via tf
- move router image into new repo (https://github.com/finboxio/rancher-lb)
- expose port 32810 for staging mongo instances
- fix treatment of host labels on register
- add target-specific terraform commands
- add staging_version output
- enable separate ami/image versioning
- fix syntax in production version definition
- enable environment-specific image versioning
- add 10m timeout before terminating stale fleet instances
- send host labels as attachments to slack on termination
- send host labels as attachments on registration
- add label io.rancher.host.external_dns_ip with public ip value on host registration
- move mongo-convoy to its own repo (https://github.com/finboxio/mongo-convoy.git)
- add production root dns records
- add analyst production infrastructure
- fix maintenance page requested url
- add finbox.io domains for maintenance and null cloudfront distributions
- fix maintenance page reload
- fix newlines for haproxy config with multiple root domains, allow 'domains' label in addition to 'domain'
- add port 32811 for secondary mongo and allow egress from host security group
- production.finbox.io dns records w/failover
- fix healthchecks
- remove production dns records
- add production infrastructure
- add maintenance/404 page to staging and dns failover
- fix healthcheck credentials file placement
- remove volume config if not mounted
- fix mysql backup for passwords with unusual characters
- use rancher hostname for backup id instead of deployment tag
- temporary fix for fleet subnet spec in us-west-2. use rancher hostname for backup id instead of deployment tag
- add mongo mms, mongo backup, and job to enforce primary status
- remove stacks from repo
- improve router error handling with proper status codes and environment-variable redirects
- add die image for testing handling of error codes returned from backends
- restore fallback url for requests with no recognized backends (handle differently from known backends with 0 servers so we can do 'not found' vs 'maintenance')
- remove mongo elb and route through default instead
- add active haproxy router
- drop ssl from mongo loadbalancer
- add mongo and convoy images to makefile
- leave curl in mongo image
- stacks on stacks of changes
- do not terminate spot instances on fleet cancellation
- add asg/fleet tags to serf members
- remove all-* checks files for out-of-service nodes
- add server fleet module
- remove server from load-balancer on termination
- separate asg-specific resources from base server infrastructure
- register with given load balancer after tagging
- remove instance from serf on termination message
- updates to status page
- move clustercheck out of cron and into persistent service
- filter out health checks to/from terminated instances
- run cron tasks as root
- add cluster status checks and s3 + cloudfront hosted web app
- add cluster status checks and s3 + cloudfront hosted web app
- only handle cluster-wide termination message if rancher is locally available
- clarify slack messages for rancher server vs rancher host
- enable configurable use of latest version in cloud config
- wait until console is ready before starting custom services
- check for failure to start rancher agent during host bootstrapping
- handle unavailability of instance metadata
- fix graceful serf exit on host termination
- use latest docker image
- handle case where instance metadata is not available to serf
- stop reporter on instance termination
- optimize packer build since we don't need our images in user-docker anymore
- reorganize source, add docker monitor to kickstart stalled user-docker, integrate aws cloudwatch events for instance termination, improved termination procedure, host and server healthcheck endpoints, run everything as system service
- updates for terraform v0.7 (data sources and string concat)
- extra attempt to remove rancher host and create a new backup on host termination
- add rancher/agent-instance to base image
- fix makefile references
- fix makefile references
- reorganize terraform, fix auto registration/deregistration and other bugs
- add rancher environment module for autoscaling host groups
- Merge pull request #1 from finboxio/blue-green
- functional blue-green server infrastructure, organized into terraform modules
- switch to elb health checks
- increase default volume size, detach mysql container from mysqlvol
- remove pid and gvwstate files from mysql backup
- add mysql-monitor to dockerfile
- start rising services at RISE_COUNT - 1
- add mysql monitor to restore primaries in case of quorum loss
- don't print an unhealthy message if monitored service is rising
- pull additional rancher images during packer build
- fix possible inconsistency in cluster size & ha launch script, pin rancher version
- improve mysql leader monitoring, improve rancher bootstrapping to avoid conflicts
- cleanup snapshot message on termination
- fix last-words config
- retry api key generation
- fix galera url concatenation
- set unhealthy status to fall count after parent failure
- make dev-tag optional for images
- fix mysqlchk when socket is not available
- fix changelog formatting
- run mysql-related processes as mysql user
- update image reference to finboxio/rancher-asg-server
- add rise/fall settings to status monitor
- ignore terraform state files
- add atlas ami reference to launch configuration
- compute quorum_size from cluster_size
- fix check for rancher auth setting when db has not been initialized
- add terraform commands, fix version retrieval, add docker image tags
- add changelog
- Initial commit