forked from Netflix/asgard
-
Notifications
You must be signed in to change notification settings - Fork 0
/
CHANGES.txt
228 lines (182 loc) · 8.34 KB
/
CHANGES.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
1.5
Features
- Platform-independent user data provider option (formatted like a properties file)
- Ability to run Asgard in one AWS account and AssumeRole to manage a different account
- Automated deployment can accept arbitrary user-chosen steps
- Each application can have an optional application group
- Each application can have optional tags
- Enable configurable maximum number of ASGs per cluster
- REST endpoints for retrieving a deployment workflow plan JSON blob, and for starting a workflow with that JSON
- Additional ways to configure Asgard to get AWS credentials, including from an IAM profile
- Updated list of AWS instance types
- Links to security groups include both name and ID
- Made it possible to skip cache loading during cache loading, using runtime flag API
- Support for HTTPS ELB listeners (Thanks Greg Dziemidowicz)
- Configurable AWS API socket timeout value
- Shared visibility all of in-memory running tasks across Asgard instances in a cluster
- Filterable instance reservations by offering type
Infrastructure
- Turn off noisy, legacy task completion emails
- Introduced AngularJS for auto deployment screens, with Grunt build and JavaScript unit tests
- Asgard's health check should fail while initializing, but then should pass forever, for desired load balancer behavior
- Asgard can register with Eureka service for conventional discoverability
- "Wither" function to have Asgard delete itself after it finishes all running in-memory tasks
- Separate SimpleDB service from Application service
- Store SWF tokens in database
- Ability to set arbitrary headers via RestClientService (Thanks e0d)
- Upgraded frigga library to allow for letters in version strings of AMI appversion tags
- Converted more JUnit tests to Spock
Bug Fixes
- List of VPCs for security group creation only includes VPCs that have conventionally labeled subnets
- List of ELBs over 400 can be listed successfully
- List of RDS DB instances over 100 can be listed successfully
- Rolling push can work with spot instances (Thanks Dale Wijnand)
- Force U.S. dollars for Amazon-listed currency amounts (Thanks Dale Wijnand)
- Allow an initial size of 0 for creating the next ASG in a cluster
1.4.2
Features
- Stop entire workflow if exception happens at any time during long auto deployment
- Auto deploy should still work even without Eureka configured
Infrastructure
- In-browser functional regression test suite with Geb
- Updated to Glisten 0.3
- Contributing guidelines document
- More unit tests
- More CodeNarc rules
Bug Fixes
- Fixed error on instance detail screen and rolling push
- Made Volume cache loading optional for initialization, since it fails often and isn't very important
1.4.1
Bug Fixes
- Support for updating load balancer in recent AWS region-accounts that have a default VPC
- Support for creating auto scaling groups with load balancers in recent AWS region-accounts that have a default VPC
- Only one auth-revoke task with many IP permissions for each security group pair change
- Zebra-stripe more lists before JavaScript loads
Infrastructure
- Some code clean up
- More unit tests
1.4
Features
- Route53 interface
- Instance sub-lists by Eureka application, where available
- Allow editing SNS Topic MessageRetentionPeriod
- Update SQS Queue policy automatically when the Queue gets subscribed to an SNS Topic
- Navigation links between associated SNS Topic and SQS Queue detail screens
- Verify ELB status for instances during auto deploy ASG health check
- Expose AWS account attributes at /account/attributes.json
- Monitor Bucket Type of "none" for applications
- Show filter on Stack list screen
Infrastructure
- Many CodeNarc rules required
- Suppress cookie warnings from remote servers in log
- Sporadic unit test concurrency-based failure fixed
- Don't pull in Glisten's transitive dependencies
- Upgrade to AWS Java SDK 1.6.6
- Update, harden, and simplify Instance Types metadata system
- Removed jsoup dependency
Bug Fixes
- Handle badly formatted Workflow Execution tags gracefully
1.3.1
Bug Fixes
- Fixed startup error for new Asgard installations that lack an existing Config.groovy file
- Fixed startup error for AWS accounts that don't yet have an Asgard SWF domain
1.3
Features
- Auto-deployment process (beta)
- Faster cache loading system initialization
- More informative startup loading screen
- Hide Desired Size input by default on ASGs that have scaling policies or scheduled actions
- Show VPC on security group
- Support for displaying notifications from Blesk
- Basic initial support for the concept of an optional Application Group field
- New, more powerful AdvancedUserDataProvider pluggable interface to replace limited legacy UserDataProvider
- Allowed Firefox read-only usage, disabled submit buttons instead of page loads (Some forms are still not FF-safe)
Infrastructure
- Integration with SWF and Glisten
- Upgraded to Grails 2.2.4
- Stop spamming log with warning about image tags
- Sped up unit test runs
- Shorter system email subject lines
- Moved filters to a package to make them easier to test
- Refactored some config access code
- Improved factory model for unit-tested instantiation of various ASG and LC related objects
- More GroovyDoc comments
- Shared IdService for generating numerical or UUID identifier strings for multiple types
Bug Fixes
- Instance types missing price info are sorted to bottom of drop-down instead of top
- Fixed a memory leak
- Prepared OneLogin integration for Java 7 and Grails 2.2.4
- Fixed bug in /instance/appversions
- Made raw instance log output null safe
- ASG force delete starts long-running task until deletion completes
- Scheduled Action links are now properly regional
- Preserve query params on login redirect from cluster screen
1.2
Features
- Choose healthiest Eureka nodes, and retry calls
- Expose ELB "other" policies, read-only for now
Infrastructure
- Upgrade to CodeNarc 0.18.1
Bug Fixes
- 404 Not Found response for non-existent DBSnapshot
- Allow two-letter app names in select2 drop-downs
- Rolling Push should retain block device mappings
- Fail scale-up operation if traffic allowed and status not UP
- New install should show correct warning in Firefox for now
1.1.2
Features
- Allow red black with no traffic and no Eureka health checks
- Use asgard user agent in AWS calls
- Add ap-southeast-2 AWS region
- Disable form submit on keyboard enter key to avoid bad ASG config creation
- Added EBS optimized to rest API for create ASG and next ASG
- Alarm dimensions other than ASG, such as Queue Name
- Enable CORS on Asgard JSON endpoints
- Prevent Firefox use until deployment GUI behavior on Firefox can be fixed
- Stack list and detail screens
Bug Fixes
- Fixes for running in Java 7
- More accurate http codes returned for anyInstance actions
- Remove dependency on brittle remote HTML parsing for hardware profiles
1.1.1
Features
- Ability to log out of OneLogin-based auth session
- Restrict GUI-based updates when using OneLogin auth
- Spot pricing for Auto Scaling Groups
- Show flags in region images
- cluster/anyInstance and autoScaling/anyInstance prefer to return a running instance
Bug Fixes
- Fix sporadic bug in ELB selects on Firefox page refresh
- Make select2 render consistently
1.1
Features
- Better select boxes
- VPC support with subnet purpose labels
- Instance types hi.4xlarge, m3.xlarge, m3.2xlarge
- ASG scheduled actions
- ASG spot pricing
- IAM instance profiles in ASGs
- ASG-ELB availability zone mismatch warnings
- SQS delivery delay
- Expose AWS SSH keys via REST
- Use Grails Standalone plugin for easier initial trials
- Configurable Eureka integration
- Ability to operate through a network proxy
- Use custom instance types before Amazon provides them
- Sortable instance table on ELB screen
Infrastructure
- grailsw script for development without an existing Grails installation
- Externalized string parsing to Frigga lib
- OneLogin integration with Shiro for authentication
- Upgrade to Grails 2.1
- Cache debugging view
Bug fixes
- Allow error-free startup on accounts without SQS
- More lenient instance types HTML parsing
- Stop task should redirect to "show"
- Show Cluster will find new ASG that is not in the cache
- Show Application with caps should not add cache duplicate
- Display remote security groups in detail without links
- Next ASG in Cluster should allow empty lists in REST API
- Rolling push should respect original SSH Keypair choice
- Universally public EBS snapshots should not be in cache