fix(provider/kubernetes): Advanced targetSize use case. #1920
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
|
||
| Validation around targetSize already happens at the Create Server Group level. This is just properly casting the | ||
| description.targetSize. | ||
| */ |
There was a problem hiding this comment.
I think this comment should just be part of the commit message. It's really only relevant with comparing the diff. You could just briefly mention the intentions of the 3 lines here.
|
|
||
| Validation around targetSize already happens at the Create Server Group level. This is just properly casting the | ||
| description.targetSize. | ||
| */ |
| description.targetSize. | ||
| */ | ||
| Integer targetSize = description.targetSize.toInteger() | ||
| if (targetSize < 0) { |
There was a problem hiding this comment.
What happens if targetSize is null? Does description.targetSize.toInteger() return -1 in that case? Just want to make sure we preserve the original behavior as well.
There was a problem hiding this comment.
targetSize cannot be null - that's part of the Server Group Validation.
| def parsedName = Names.parseName(replicaSetName) | ||
| def targetSize = description.targetSize ?: description.capacity?.desired | ||
| if (targetSize < 0) { |
Z3r0Sum
force-pushed
the
fix-replicaset-creation
branch
from
6780331 to
80eb0fd
Compare
| value of 1, which will happen if the targetSize is represented by a null value at this juncture in the code. | ||
| */ | ||
| Integer targetSize = description.targetSize.toInteger() | ||
| if (targetSize < 0) { |
There was a problem hiding this comment.
Since targetSize can never be null or < 0, I don't think description.capacity will ever be considered.
There was a problem hiding this comment.
Yeah, I'll need to reason about on this... this will need to change.
There was a problem hiding this comment.
I'm thinking the right way to approach this would be the following:
Defer to the desired source capacity, so if the targetSize is less than it - the desired source capacity wins. This still covers the use case of this PR as you wouldn't set a desired source capacity of > 0 if you wanted a targetSize of 0 or you wanted the resource at 0. Essentially, desired source capacity always trumps targetSize. Thoughts?
Z3r0Sum
force-pushed
the
fix-replicaset-creation
branch
from
80eb0fd to
6c3e42b
Compare
|
Scenarios Tested:
Should be noted --- it appears desired capacity sets the targetSize
Edit: Interesting based on the UI interaction, it would seem desired capacity already overrides targetSize. |
|
@lwander I'd appreciate you weighing in on this as you probably have some opinions on what should get precedence... |
Z3r0Sum
force-pushed
the
fix-replicaset-creation
branch
2 times, most recently
from
de7926d to
5f5caf8
Compare
|
It seems like the desired source capacity is always null and just sets the targetSize, is this assumption correct? |
Z3r0Sum
force-pushed
the
fix-replicaset-creation
branch
from
5f5caf8 to
22b2478
Compare
lwander
reviewed
Sep 25, 2017
| def targetSize | ||
| if (description.targetSize == 0) { | ||
| targetSize = description.targetSize | ||
| } else { targetSize = description.targetSize ?: description.capacity?.desired } |
There was a problem hiding this comment.
Could you put this onto a newline?
lwander
reviewed
Sep 25, 2017
| def targetSize | ||
| if (description.targetSize == 0) { | ||
| targetSize = description.targetSize | ||
| } else { targetSize = description.targetSize ?: description.capacity?.desired } |
lwander
approved these changes
Sep 25, 2017
Z3r0Sum
force-pushed
the
fix-replicaset-creation
branch
from
22b2478 to
f54ce05
Compare
The elvis operator before was not evaluating truthiness correctly for a value of 0. A User could and may want to deploy an update ReplicaSet of 0 and Scale it later. Or perhaps it's bundled in a Spinnaker Pipeline with multiple Server Groups and one of the Server Groups is scaled to 0. If the User is implementing 'useSourceCapacity' we would want that 0 value respected and not overwritten with a value of 1. The proposal is that the desired capacity trumps targetSize. It's reasonable to assume that someone who wants a targetSize of 0 would also have a desired capacity of 0. However, we shouldn't use the targetSize if the desired capacity is greater than its value.
Z3r0Sum
force-pushed
the
fix-replicaset-creation
branch
from
f54ce05 to
56ade88
Compare
ethanfrogers
approved these changes
Sep 25, 2017
BrunoCarrier
added a commit
to lookout/clouddriver
that referenced
this pull request
Oct 26, 2017
* feat(provider/kubernetes): Add spectator to v2 client (spinnaker#1866) * refactor(provider/kubernetes): Share caching agent abstraction (spinnaker#1865) * feat(provider/kubernetes): Fix on demand caching (spinnaker#1870) * chore(provider/kubernetes): On demand caching tests (spinnaker#1871) * feat(provider/kuberenetes): V2 deployments (spinnaker#1868) * feat(provider/kubernetes): Cache arbitrary resource (spinnaker#1873) * feat(search): update application search (spinnaker#1875) * add ability to pass search filters down to application searching in front50. * Additional fix for issue 1632, not able to find AMI with encrypted snapshots in target/managed account (spinnaker#1846) * docs(intellij): Describe how to enable lombok in intellij (spinnaker#1869) I figured we'd add a quick one-line how-to for getting gradle to build your intellij project too. * feat(provider/kubernetes): Infer relationships from ownerReference (spinnaker#1876) * feat(provider/kubernetes): Infer relationships from ownerReference * feat(provider/kubernetes): Cache ancillary resources * feat(search): add filters to project search (spinnaker#1878) * feat(provider/ecs): Made EcsCloudProvider public (spinnaker#1848) * feat(provider/kubernetes): Cache pods (spinnaker#1879) * feat(provider/kubernetes): Annotate templates as well (spinnaker#1880) A number of k8s resources use a spec.template to deploy a managed resource. These need to be annotated by Spinnaker as well. * fix(aws/tags): ensure tag sync is always performed during allow launch * fix(search): allow SearchProvider filter exclusion (spinnaker#1883) * allows search providers to specify an exclusion list for filter keys to optionally be removed before executing a search. * feat(provider/aws): Remove ability to automatically remove dependencies when deleting a security group (spinnaker#1894) * fix(provider/aws): trim whitespace when newlines are detected (spinnaker#1844) * fix(google): Add statusCode tag to google API metrics (spinnaker#1886) This also refactors all the tests involving google API metrics. * fix(provider/kubernetes): Add missing credential props (spinnaker#1897) * feat(provider/kubernetes): add daemon & stateful set support * fix(provider/aws): ensure STSAssumeRoleSessionCredentialsProvider get the right endpoint (spinnaker#1888) * feat(provider/kubernetes): v2 generic on demand caching (spinnaker#1884) * refactor(provider/kubernetes): v2 include api version in resource id (spinnaker#1901) * feat(provider/docker): add catalogFile option (spinnaker#1890) adds the `catalogFile` option to the Docker Registry provider. This is an attempt to make it easier to use registries that don't support `v2/_catalog` (which happens to be a lot). This file can be managed by an external process. Each caching cycle, Clouddriver reads this file to determine the list of repositories it should cache tags for. the format of the `configFile` is: ``` { "repositories":[ "library/nginx", "library/busybox" ] } ``` * fix(provider/docker): fix client constructor (spinnaker#1903) Bug introduced in `f70d6ea`. I updated the constructor to include `catalogFile` but didn't update the constructor for those using `password` and `passwordFile` configurations. This fixes that potential issue. * feat(provider/kubernetes): v2 Cache lb & scg relationships (spinnaker#1902) * Revert " fix(provider/aws): ensure STSAssumeRoleSessionCredentialsProvider get the right endpoint (spinnaker#1888)" (spinnaker#1904) This reverts commit 828fbb7. * feat(provider/kubernetes): v2 version deployed manifests as resources (spinnaker#1905) * refactor(provider/kubernetes): V2 swap kind & version in cache key (spinnaker#1907) For consistency with naming & other resources. * fix(provider/aws): Clone spot price if not explicitly specified. (spinnaker#1908) * refactor(provider/kubernetes): Segregate k8s by provider version (spinnaker#1909) * feat(provider/kubernetes): Add proper support for source capacity. (spinnaker#1906) Fully addresses: spinnaker/spinnaker#1657 Also accounts for the possibility that the user might want to set their own targetSize while ignoring the source capacity. Deck work is still needed to expose "useSourceCapacity" in the UI. Fix cross namespace server groups * feat(provider/kubernetes): Register k8s moniker for manifests (spinnaker#1910) * feat(provider/kubernetes): V2 cluster provider (spinnaker#1911) * refactor(core): s/version/providerVersion (spinnaker#1914) Having the account property named 'version' was confusing the concept of the version the provider was at, with the version the cloud environment was at. * refactor(provider/kubernetes): Rename duplicate classes (spinnaker#1913) * feat(moniker): Adds moniker to all clusters & server groups (spinnaker#1915) * chore(javadocs): Fix lt usage (spinnaker#1916) * feat(provider/kubernetes): Attach k8s annotation monikers to v2 (spinnaker#1917) resources * fix(aws): CopyLastAsgAtomicOperation instance monitoring. When not explicitly specified, instance monitoring would default to true due to groovy truth casting a non null object to true instead of traversing into its enabled attribute * fix(web): Guard against not finding a ClusterProvider (spinnaker#1921) * feat(search): add fallback query param to search (spinnaker#1912) * in an effort to speed up searching, this adds logic that attempts to pick a "good" value to use as a search query string rather than just using the empty string and pulling all the keys for a given type. * feat(provider/kubernetes): V2 Cluster details (spinnaker#1918) * chore(provider/kubernetes): Use version 0.2.0 of client-java (spinnaker#1923) * fix(provider/kubernetes): Advanced targetSize use case. (spinnaker#1920) The elvis operator before was not evaluating truthiness correctly for a value of 0. A User could and may want to deploy an update ReplicaSet of 0 and Scale it later. Or perhaps it's bundled in a Spinnaker Pipeline with multiple Server Groups and one of the Server Groups is scaled to 0. If the User is implementing 'useSourceCapacity' we would want that 0 value respected and not overwritten with a value of 1. The proposal is that the desired capacity trumps targetSize. It's reasonable to assume that someone who wants a targetSize of 0 would also have a desired capacity of 0. However, we shouldn't use the targetSize if the desired capacity is greater than its value. * refactor(provider/kubernetes): v2 cache view lives in cache package (spinnaker#1926) * feat(provider/kubernetes): v2 cache network policies (spinnaker#1927) * feat(provider/kubernetes): Build v2 creds based on context & kubeconfigFile (spinnaker#1925) * feat(provider/kubernetes): Integrate moniker into v2 (spinnaker#1929) * fix(provider/kubernetes): v2 guard against empty cache relationships (spinnaker#1932) * fix(provider/kuberentes): v2 guard against non-existent clusters (spinnaker#1934) * feat(moniker): adds moniker to server group view model (spinnaker#1936) * feat(provider/kubernetes): Cache v2 resources as artifacts (spinnaker#1931) * fix(all): Broken k8s server group lookup breaking all components (spinnaker#1937) @lwander I'm merging for short term fix since everyone is affected. However see my question about returning null and whether there are other implications. * feat(provider/google): Adds support for multiple named ports in load balancers. (spinnaker#1930) * refactor(provider/kubernetes): refactor artifact gen (spinnaker#1938) * feat(moniker): Adds moniker to pendingOnDemandResults (spinnaker#1940) * fix(provider/google): Fix named port defaults in server group deploy. (spinnaker#1939) * feat(moniker): Add monikers to security groups & load balancers (spinnaker#1941) * feat(provider/google): Support connection draining for LBs (spinnaker#1943) * fix(provider/gce): s/it/cacheData/ in zonal svg caching agent. (spinnaker#1944) * feat(provider/docker): Added insecure registry support (spinnaker#1887) * feat(provider/docker): Added insecure registry support Added an insecureRegistry flag to clouddriver-docker registry support that will skip SSL certificate validation. This will allow for addressing spinnaker/halyard#680. * Update DockerRegistryNamedAccountCredentials.groovy * Update DockerRegistryClientSpec.groovy * feat(appengine): Deploy from Google Cloud Storage [accounts]. (spinnaker#1935) * fix(provider/docker): Assign proper Spinnaker/<version> user agent (spinnaker#1946) * feat(provider/kubernetes): Properly version all versioned resources (spinnaker#1945) * fix(rollbacks): support for tolerating some instance failures (spinnaker#1922) * fix(provider/gce): Paginate server groups when calculating next name. (spinnaker#1948) * fix(appengine): Dont assume storage.gce.enabled=true (spinnaker#1953) * fix(provider/kubernetes): fix k8s client configuration and image id p… (spinnaker#1933) * fix(provider/kubernetes): fix k8s client configuration and image id problem. * (feat/cats) Add support for an error interval for clustered agent schedulers (spinnaker#1895) * feat(search): cull missing ASGs before returning (spinnaker#1951) * removes ASGs that show up in the search results from the redis cache but don't actually exist. * fix(provider/kubernetes): Fixed an issue with jobProvider failing out other providers. (spinnaker#1690) * fix(provider/dcos): fix cluster provider to guard against errors from invalid ids (spinnaker#1928) * feat(core): Add provider version to creds controller (spinnaker#1958) * feat(appengine): Add AppengineStorageController to list accounts (spinnaker#1957) * chore(provider/kubernetes): Bump client to 1.0.0-beta1 (spinnaker#1961) * fix(provider/kubernetes): v1 client call to list pods (spinnaker#1963) * fix(provider/aws): Return load balancers for all target groups associated with an application (spinnaker#1955) * refactor(provider/kubernetes): V2 deploy 1 manifest at a time (spinnaker#1965) * feat(provider/gae): Modify deploy description and validators for artifacts. (spinnaker#1966) * feat(provider/kubernetes): Feature-flag debug mode (spinnaker#1968) * fix(provider/kubernetes): Don't store-unowned artifacts (spinnaker#1969) * feat(provider/kuberentes): Application provider (spinnaker#1971) * fix(provider/aws): fix issue with VPCZoneIdentifier containing deleted subnets (spinnaker#1972) * chore(provider/docker): Only log trailing whitespace warning during failure (spinnaker#1973) * feat(provider/kubernetes): V2 cluster provider finishing touches (spinnaker#1974) * fix(provider/aws): Allow updating healthy threshold count on target groups (spinnaker#1977) * feat(moniker): Add getServerGroup endpoint that does not require application name. (spinnaker#1967) * feat(provider/kubernetes): v2 instance counts (spinnaker#1975) * feat(provider/kuberentes): Register kind mapping using autowired deploy (spinnaker#1978) * refactor(provider/kubernetes): Move manifest desc. into package (spinnaker#1979) * refactor(provider/kubernetes): Move ops w/ kind into v1 only (spinnaker#1980) * feat(provider/kubernetes): Support resize statefulset operation. (spinnaker#1976) * feat(dynomite): hashtags & dynomite pipelines (spinnaker#1832) * refactor(provider/google): Rely on cached instance details instead of calling platform MIG.listInstances() api. (spinnaker#1983) * feat(provider/amazon): Create app elb specific security group for new ELBs (spinnaker#1982) - Autocreate elb security group and ingress on application security group * feat(core/search): Improve search speed: faster validation that search keys exists (spinnaker#1984) * feat(provider/kubernetes): Support resize operations (spinnaker#1981) * feat(provider/kubernetes): V2 instance state (spinnaker#1986) * fix(provider/kubernetes): Fix v2 operation converter (spinnaker#1987) * feat(provider/gce): Replace instanceTemplate get with list call. (spinnaker#1990) * refactor(provider/kubernetes): use deployed names in result (spinnaker#1989) * fix(core): Support the eviction of stale cache key identifiers (spinnaker#1985) This PR attempts to clean up situations where the `:members` set for a given cached type contains keys that no longer exist. In particular, a `CachingAgent` can optionally return a pattern glob for each authoritative type. Any `:members` key matching the pattern glob BUT not included in the most recent caching cycle will be evicted. This has been enabled on a select few `aws` caching agents. * fix(aws): Do not do stale key evictions of amazon load balancers (spinnaker#1993) Seeing some mixed results when application/network and classic load balancers are being used. * feat(provider/kubernetes): v2 on-demand manifest caching (spinnaker#1988) * feat(provider/aws): add executionId to user-agent for cloudtrail (spinnaker#1992) * fix(aws): Enable stale key evictions for amazon load balancers (spinnaker#1994) Key structures vary between classic and application/networking load balancers. * feat(provider/kubernetes): Re-encode api types (spinnaker#1991) * feat(logging) compile logstashEncoder for json logging (spinnaker#1995) * fix(provider/gae): Set path correct when application root non-empty. (spinnaker#1970) * chore(dependencies): Bump spinnaker-dependencies to 0.115.0. (spinnaker#1996) * fix hidden exception while pulling docker tags (spinnaker#1708) * fix(provider/kubernetes): reads correct fields from cache refresh (spinnaker#1998) * feat(provider/kubernetes): v2 delete operation (spinnaker#1999) * feat(provider/dcos): Adding proxy for marathon-client to make use of spectator. (spinnaker#1960) * fix(appengine): fix deploy failures (spinnaker#2001) * feat(provider/kubernetes): Supply app name in attributes (spinnaker#2003) * feat(provider/kuberentes): implement capacity (spinnaker#2004) * feat(cats): Adding cache gzip compression (spinnaker#2000) * feat(provider/kubernetes): Update deployable's patch behavior (spinnaker#2006) * feat(provider/kubernetes): Show more in the error message (spinnaker#2007) * feat(provider/kubernetes): V2 Cache services (spinnaker#2009) * fix(web): Fixing missing property exception (spinnaker#2008) * fix(cats): Get bytes in consistent chatset during decompress (spinnaker#2010) * fix(cats): Remove use of unmodifiable collections (spinnaker#2011) * feat(provider/kubernetes): cache non-obvious relationship (spinnaker#2012) * feat(provider/kubernetes): load balancer -> sg relationship (spinnaker#2013) * refactor(aws): Simplified ELB security group autocreation (spinnaker#1997) - updated code that allows ingress from elb group to app security group * fix(provider/gae): Delete gcs target directory in deploy. (spinnaker#2015) * fix(provider/gae): Delete gcs target directory in deploy. * fix(storage): Change the storage account endpoint to '/storage'. * feat(provider/kubernetes): Enable setting of labels on k8s Service objects spinnaker/spinnaker#2035 (spinnaker#2014) * fix(provider/kubernetes): allows dots in secretNames (spinnaker#2017) The kubernetes api allows dots in secret names: https://github.com/kubernetes/kubernetes/blob/e5ac41331119bb8c67004b8efaa0c3d942b31d3a/pkg/api/validation/validation.go#L160-L162 https://github.com/kubernetes/kubernetes/blob/8d7d7a5e0d4d7e75f5a860574346944b8cc0fc43/staging/src/k8s.io/apimachinery/pkg/util/validation/validation.go#L126-L142 https://github.com/kubernetes/kubernetes/blob/e5ac41331119bb8c67004b8efaa0c3d942b31d3a/pkg/api/validation/validation.go#L1488 This allows dots in the secretName field of volume source definitions. * perf(cluster): Avoid prematurely fetching full server groups (spinnaker#2019) This is an optimization for fetching target server groups by `CURRENT`, `PREVIOUS`, `OLDEST` and `FAIL`. For these strategies, the unexpanded server group is sufficient to narrow down to a match. Once a match is found, _only_ that server group will be expanded. This only works for cluster providers that support loading both expanded and unexpanded server groups (ie. aws). The behavior for other providers should be unchanged. * fix(aws): Fix NPE in elb group autocreation (spinnaker#2020) - fixed NPE due to missing credentials in security group description request * feat(provider/kubernetes): Enable annotations, labels and secrets for security groups (aka ingress resources) (spinnaker#2000) (spinnaker#2005) * feat(google): Instrument individual google API calls. (spinnaker#2016) * fix(cats): Remove unmodifiable collections from modifiable codepaths (spinnaker#2022) * fix(provider/gce): Fix broken deploys with namedPorts. (spinnaker#2023) * fix(aws): Only autocreate app elb security group on Create (spinnaker#2025) - Updated to only auto create <appname>-elb group for new ELBs * feat(provider/aws): Support specifying explicit subnet ids for deploy (spinnaker#2026) Expectation is that a `subnetType` will still be provided and that `subnetIds` will be a subnet of those valid for `subnetType`. When explicit subnet ids are provided (or inherited), the newly created server group will be tagged with: `SPINNAKER_SUBNET_ID_OVERRIDE`: `"comma-separated list of subnet ids"` * fix(provider/gce): Paginate instance template list in svg caching. (spinnaker#2027) * fix(eureka): fix instanceId on eureka caching The instanceId was intended to be the value in dataCenterInfo.metadata but due to some Jackson behaviour the top level instanceId from the discovery registration was written over the dataCenterInfo.metadata during deserialization. This changes to explicitly handle the top level instanceId value in the JsonCreator factory method to prevent the unhandled property behaviour from kicking in. * Merged upstream, updated AbstractEcsCachingAgent and Keys.java
BrunoCarrier
added a commit
to lookout/clouddriver
that referenced
this pull request
Oct 26, 2017
* feat(provider/kubernetes): Add spectator to v2 client (spinnaker#1866) * refactor(provider/kubernetes): Share caching agent abstraction (spinnaker#1865) * feat(provider/kubernetes): Fix on demand caching (spinnaker#1870) * chore(provider/kubernetes): On demand caching tests (spinnaker#1871) * feat(provider/kuberenetes): V2 deployments (spinnaker#1868) * feat(provider/kubernetes): Cache arbitrary resource (spinnaker#1873) * feat(search): update application search (spinnaker#1875) * add ability to pass search filters down to application searching in front50. * Additional fix for issue 1632, not able to find AMI with encrypted snapshots in target/managed account (spinnaker#1846) * docs(intellij): Describe how to enable lombok in intellij (spinnaker#1869) I figured we'd add a quick one-line how-to for getting gradle to build your intellij project too. * feat(provider/kubernetes): Infer relationships from ownerReference (spinnaker#1876) * feat(provider/kubernetes): Infer relationships from ownerReference * feat(provider/kubernetes): Cache ancillary resources * feat(search): add filters to project search (spinnaker#1878) * feat(provider/ecs): Made EcsCloudProvider public (spinnaker#1848) * feat(provider/kubernetes): Cache pods (spinnaker#1879) * feat(provider/kubernetes): Annotate templates as well (spinnaker#1880) A number of k8s resources use a spec.template to deploy a managed resource. These need to be annotated by Spinnaker as well. * fix(aws/tags): ensure tag sync is always performed during allow launch * fix(search): allow SearchProvider filter exclusion (spinnaker#1883) * allows search providers to specify an exclusion list for filter keys to optionally be removed before executing a search. * feat(provider/aws): Remove ability to automatically remove dependencies when deleting a security group (spinnaker#1894) * fix(provider/aws): trim whitespace when newlines are detected (spinnaker#1844) * fix(google): Add statusCode tag to google API metrics (spinnaker#1886) This also refactors all the tests involving google API metrics. * fix(provider/kubernetes): Add missing credential props (spinnaker#1897) * feat(provider/kubernetes): add daemon & stateful set support * fix(provider/aws): ensure STSAssumeRoleSessionCredentialsProvider get the right endpoint (spinnaker#1888) * feat(provider/kubernetes): v2 generic on demand caching (spinnaker#1884) * refactor(provider/kubernetes): v2 include api version in resource id (spinnaker#1901) * feat(provider/docker): add catalogFile option (spinnaker#1890) adds the `catalogFile` option to the Docker Registry provider. This is an attempt to make it easier to use registries that don't support `v2/_catalog` (which happens to be a lot). This file can be managed by an external process. Each caching cycle, Clouddriver reads this file to determine the list of repositories it should cache tags for. the format of the `configFile` is: ``` { "repositories":[ "library/nginx", "library/busybox" ] } ``` * fix(provider/docker): fix client constructor (spinnaker#1903) Bug introduced in `f70d6ea`. I updated the constructor to include `catalogFile` but didn't update the constructor for those using `password` and `passwordFile` configurations. This fixes that potential issue. * feat(provider/kubernetes): v2 Cache lb & scg relationships (spinnaker#1902) * Revert " fix(provider/aws): ensure STSAssumeRoleSessionCredentialsProvider get the right endpoint (spinnaker#1888)" (spinnaker#1904) This reverts commit 828fbb7. * feat(provider/kubernetes): v2 version deployed manifests as resources (spinnaker#1905) * refactor(provider/kubernetes): V2 swap kind & version in cache key (spinnaker#1907) For consistency with naming & other resources. * fix(provider/aws): Clone spot price if not explicitly specified. (spinnaker#1908) * refactor(provider/kubernetes): Segregate k8s by provider version (spinnaker#1909) * feat(provider/kubernetes): Add proper support for source capacity. (spinnaker#1906) Fully addresses: spinnaker/spinnaker#1657 Also accounts for the possibility that the user might want to set their own targetSize while ignoring the source capacity. Deck work is still needed to expose "useSourceCapacity" in the UI. Fix cross namespace server groups * feat(provider/kubernetes): Register k8s moniker for manifests (spinnaker#1910) * feat(provider/kubernetes): V2 cluster provider (spinnaker#1911) * refactor(core): s/version/providerVersion (spinnaker#1914) Having the account property named 'version' was confusing the concept of the version the provider was at, with the version the cloud environment was at. * refactor(provider/kubernetes): Rename duplicate classes (spinnaker#1913) * feat(moniker): Adds moniker to all clusters & server groups (spinnaker#1915) * chore(javadocs): Fix lt usage (spinnaker#1916) * feat(provider/kubernetes): Attach k8s annotation monikers to v2 (spinnaker#1917) resources * fix(aws): CopyLastAsgAtomicOperation instance monitoring. When not explicitly specified, instance monitoring would default to true due to groovy truth casting a non null object to true instead of traversing into its enabled attribute * fix(web): Guard against not finding a ClusterProvider (spinnaker#1921) * feat(search): add fallback query param to search (spinnaker#1912) * in an effort to speed up searching, this adds logic that attempts to pick a "good" value to use as a search query string rather than just using the empty string and pulling all the keys for a given type. * feat(provider/kubernetes): V2 Cluster details (spinnaker#1918) * chore(provider/kubernetes): Use version 0.2.0 of client-java (spinnaker#1923) * fix(provider/kubernetes): Advanced targetSize use case. (spinnaker#1920) The elvis operator before was not evaluating truthiness correctly for a value of 0. A User could and may want to deploy an update ReplicaSet of 0 and Scale it later. Or perhaps it's bundled in a Spinnaker Pipeline with multiple Server Groups and one of the Server Groups is scaled to 0. If the User is implementing 'useSourceCapacity' we would want that 0 value respected and not overwritten with a value of 1. The proposal is that the desired capacity trumps targetSize. It's reasonable to assume that someone who wants a targetSize of 0 would also have a desired capacity of 0. However, we shouldn't use the targetSize if the desired capacity is greater than its value. * refactor(provider/kubernetes): v2 cache view lives in cache package (spinnaker#1926) * feat(provider/kubernetes): v2 cache network policies (spinnaker#1927) * feat(provider/kubernetes): Build v2 creds based on context & kubeconfigFile (spinnaker#1925) * feat(provider/kubernetes): Integrate moniker into v2 (spinnaker#1929) * fix(provider/kubernetes): v2 guard against empty cache relationships (spinnaker#1932) * fix(provider/kuberentes): v2 guard against non-existent clusters (spinnaker#1934) * feat(moniker): adds moniker to server group view model (spinnaker#1936) * feat(provider/kubernetes): Cache v2 resources as artifacts (spinnaker#1931) * fix(all): Broken k8s server group lookup breaking all components (spinnaker#1937) @lwander I'm merging for short term fix since everyone is affected. However see my question about returning null and whether there are other implications. * feat(provider/google): Adds support for multiple named ports in load balancers. (spinnaker#1930) * refactor(provider/kubernetes): refactor artifact gen (spinnaker#1938) * feat(moniker): Adds moniker to pendingOnDemandResults (spinnaker#1940) * fix(provider/google): Fix named port defaults in server group deploy. (spinnaker#1939) * feat(moniker): Add monikers to security groups & load balancers (spinnaker#1941) * feat(provider/google): Support connection draining for LBs (spinnaker#1943) * fix(provider/gce): s/it/cacheData/ in zonal svg caching agent. (spinnaker#1944) * feat(provider/docker): Added insecure registry support (spinnaker#1887) * feat(provider/docker): Added insecure registry support Added an insecureRegistry flag to clouddriver-docker registry support that will skip SSL certificate validation. This will allow for addressing spinnaker/halyard#680. * Update DockerRegistryNamedAccountCredentials.groovy * Update DockerRegistryClientSpec.groovy * feat(appengine): Deploy from Google Cloud Storage [accounts]. (spinnaker#1935) * fix(provider/docker): Assign proper Spinnaker/<version> user agent (spinnaker#1946) * feat(provider/kubernetes): Properly version all versioned resources (spinnaker#1945) * fix(rollbacks): support for tolerating some instance failures (spinnaker#1922) * fix(provider/gce): Paginate server groups when calculating next name. (spinnaker#1948) * fix(appengine): Dont assume storage.gce.enabled=true (spinnaker#1953) * fix(provider/kubernetes): fix k8s client configuration and image id p… (spinnaker#1933) * fix(provider/kubernetes): fix k8s client configuration and image id problem. * (feat/cats) Add support for an error interval for clustered agent schedulers (spinnaker#1895) * feat(search): cull missing ASGs before returning (spinnaker#1951) * removes ASGs that show up in the search results from the redis cache but don't actually exist. * fix(provider/kubernetes): Fixed an issue with jobProvider failing out other providers. (spinnaker#1690) * fix(provider/dcos): fix cluster provider to guard against errors from invalid ids (spinnaker#1928) * feat(core): Add provider version to creds controller (spinnaker#1958) * feat(appengine): Add AppengineStorageController to list accounts (spinnaker#1957) * chore(provider/kubernetes): Bump client to 1.0.0-beta1 (spinnaker#1961) * fix(provider/kubernetes): v1 client call to list pods (spinnaker#1963) * fix(provider/aws): Return load balancers for all target groups associated with an application (spinnaker#1955) * refactor(provider/kubernetes): V2 deploy 1 manifest at a time (spinnaker#1965) * feat(provider/gae): Modify deploy description and validators for artifacts. (spinnaker#1966) * feat(provider/kubernetes): Feature-flag debug mode (spinnaker#1968) * fix(provider/kubernetes): Don't store-unowned artifacts (spinnaker#1969) * feat(provider/kuberentes): Application provider (spinnaker#1971) * fix(provider/aws): fix issue with VPCZoneIdentifier containing deleted subnets (spinnaker#1972) * chore(provider/docker): Only log trailing whitespace warning during failure (spinnaker#1973) * feat(provider/kubernetes): V2 cluster provider finishing touches (spinnaker#1974) * fix(provider/aws): Allow updating healthy threshold count on target groups (spinnaker#1977) * feat(moniker): Add getServerGroup endpoint that does not require application name. (spinnaker#1967) * feat(provider/kubernetes): v2 instance counts (spinnaker#1975) * feat(provider/kuberentes): Register kind mapping using autowired deploy (spinnaker#1978) * refactor(provider/kubernetes): Move manifest desc. into package (spinnaker#1979) * refactor(provider/kubernetes): Move ops w/ kind into v1 only (spinnaker#1980) * feat(provider/kubernetes): Support resize statefulset operation. (spinnaker#1976) * feat(dynomite): hashtags & dynomite pipelines (spinnaker#1832) * refactor(provider/google): Rely on cached instance details instead of calling platform MIG.listInstances() api. (spinnaker#1983) * feat(provider/amazon): Create app elb specific security group for new ELBs (spinnaker#1982) - Autocreate elb security group and ingress on application security group * feat(core/search): Improve search speed: faster validation that search keys exists (spinnaker#1984) * feat(provider/kubernetes): Support resize operations (spinnaker#1981) * feat(provider/kubernetes): V2 instance state (spinnaker#1986) * fix(provider/kubernetes): Fix v2 operation converter (spinnaker#1987) * feat(provider/gce): Replace instanceTemplate get with list call. (spinnaker#1990) * refactor(provider/kubernetes): use deployed names in result (spinnaker#1989) * fix(core): Support the eviction of stale cache key identifiers (spinnaker#1985) This PR attempts to clean up situations where the `:members` set for a given cached type contains keys that no longer exist. In particular, a `CachingAgent` can optionally return a pattern glob for each authoritative type. Any `:members` key matching the pattern glob BUT not included in the most recent caching cycle will be evicted. This has been enabled on a select few `aws` caching agents. * fix(aws): Do not do stale key evictions of amazon load balancers (spinnaker#1993) Seeing some mixed results when application/network and classic load balancers are being used. * feat(provider/kubernetes): v2 on-demand manifest caching (spinnaker#1988) * feat(provider/aws): add executionId to user-agent for cloudtrail (spinnaker#1992) * fix(aws): Enable stale key evictions for amazon load balancers (spinnaker#1994) Key structures vary between classic and application/networking load balancers. * feat(provider/kubernetes): Re-encode api types (spinnaker#1991) * feat(logging) compile logstashEncoder for json logging (spinnaker#1995) * fix(provider/gae): Set path correct when application root non-empty. (spinnaker#1970) * chore(dependencies): Bump spinnaker-dependencies to 0.115.0. (spinnaker#1996) * fix hidden exception while pulling docker tags (spinnaker#1708) * fix(provider/kubernetes): reads correct fields from cache refresh (spinnaker#1998) * feat(provider/kubernetes): v2 delete operation (spinnaker#1999) * feat(provider/dcos): Adding proxy for marathon-client to make use of spectator. (spinnaker#1960) * fix(appengine): fix deploy failures (spinnaker#2001) * feat(provider/kubernetes): Supply app name in attributes (spinnaker#2003) * feat(provider/kuberentes): implement capacity (spinnaker#2004) * feat(cats): Adding cache gzip compression (spinnaker#2000) * feat(provider/kubernetes): Update deployable's patch behavior (spinnaker#2006) * feat(provider/kubernetes): Show more in the error message (spinnaker#2007) * feat(provider/kubernetes): V2 Cache services (spinnaker#2009) * fix(web): Fixing missing property exception (spinnaker#2008) * fix(cats): Get bytes in consistent chatset during decompress (spinnaker#2010) * fix(cats): Remove use of unmodifiable collections (spinnaker#2011) * feat(provider/kubernetes): cache non-obvious relationship (spinnaker#2012) * feat(provider/kubernetes): load balancer -> sg relationship (spinnaker#2013) * refactor(aws): Simplified ELB security group autocreation (spinnaker#1997) - updated code that allows ingress from elb group to app security group * fix(provider/gae): Delete gcs target directory in deploy. (spinnaker#2015) * fix(provider/gae): Delete gcs target directory in deploy. * fix(storage): Change the storage account endpoint to '/storage'. * feat(provider/kubernetes): Enable setting of labels on k8s Service objects spinnaker/spinnaker#2035 (spinnaker#2014) * fix(provider/kubernetes): allows dots in secretNames (spinnaker#2017) The kubernetes api allows dots in secret names: https://github.com/kubernetes/kubernetes/blob/e5ac41331119bb8c67004b8efaa0c3d942b31d3a/pkg/api/validation/validation.go#L160-L162 https://github.com/kubernetes/kubernetes/blob/8d7d7a5e0d4d7e75f5a860574346944b8cc0fc43/staging/src/k8s.io/apimachinery/pkg/util/validation/validation.go#L126-L142 https://github.com/kubernetes/kubernetes/blob/e5ac41331119bb8c67004b8efaa0c3d942b31d3a/pkg/api/validation/validation.go#L1488 This allows dots in the secretName field of volume source definitions. * perf(cluster): Avoid prematurely fetching full server groups (spinnaker#2019) This is an optimization for fetching target server groups by `CURRENT`, `PREVIOUS`, `OLDEST` and `FAIL`. For these strategies, the unexpanded server group is sufficient to narrow down to a match. Once a match is found, _only_ that server group will be expanded. This only works for cluster providers that support loading both expanded and unexpanded server groups (ie. aws). The behavior for other providers should be unchanged. * fix(aws): Fix NPE in elb group autocreation (spinnaker#2020) - fixed NPE due to missing credentials in security group description request * feat(provider/kubernetes): Enable annotations, labels and secrets for security groups (aka ingress resources) (spinnaker#2000) (spinnaker#2005) * feat(google): Instrument individual google API calls. (spinnaker#2016) * fix(cats): Remove unmodifiable collections from modifiable codepaths (spinnaker#2022) * fix(provider/gce): Fix broken deploys with namedPorts. (spinnaker#2023) * fix(aws): Only autocreate app elb security group on Create (spinnaker#2025) - Updated to only auto create <appname>-elb group for new ELBs * feat(provider/aws): Support specifying explicit subnet ids for deploy (spinnaker#2026) Expectation is that a `subnetType` will still be provided and that `subnetIds` will be a subnet of those valid for `subnetType`. When explicit subnet ids are provided (or inherited), the newly created server group will be tagged with: `SPINNAKER_SUBNET_ID_OVERRIDE`: `"comma-separated list of subnet ids"` * fix(provider/gce): Paginate instance template list in svg caching. (spinnaker#2027) * fix(eureka): fix instanceId on eureka caching The instanceId was intended to be the value in dataCenterInfo.metadata but due to some Jackson behaviour the top level instanceId from the discovery registration was written over the dataCenterInfo.metadata during deserialization. This changes to explicitly handle the top level instanceId value in the JsonCreator factory method to prevent the unhandled property behaviour from kicking in. * target group is now a string, instead of an array of strigns
BrunoCarrier
added a commit
to lookout/clouddriver
that referenced
this pull request
Nov 6, 2017
* feat(provider/kubernetes): Add spectator to v2 client (spinnaker#1866) * refactor(provider/kubernetes): Share caching agent abstraction (spinnaker#1865) * feat(provider/kubernetes): Fix on demand caching (spinnaker#1870) * chore(provider/kubernetes): On demand caching tests (spinnaker#1871) * feat(provider/kuberenetes): V2 deployments (spinnaker#1868) * feat(provider/kubernetes): Cache arbitrary resource (spinnaker#1873) * feat(search): update application search (spinnaker#1875) * add ability to pass search filters down to application searching in front50. * Additional fix for issue 1632, not able to find AMI with encrypted snapshots in target/managed account (spinnaker#1846) * docs(intellij): Describe how to enable lombok in intellij (spinnaker#1869) I figured we'd add a quick one-line how-to for getting gradle to build your intellij project too. * feat(provider/kubernetes): Infer relationships from ownerReference (spinnaker#1876) * feat(provider/kubernetes): Infer relationships from ownerReference * feat(provider/kubernetes): Cache ancillary resources * feat(search): add filters to project search (spinnaker#1878) * feat(provider/ecs): Made EcsCloudProvider public (spinnaker#1848) * feat(provider/kubernetes): Cache pods (spinnaker#1879) * feat(provider/kubernetes): Annotate templates as well (spinnaker#1880) A number of k8s resources use a spec.template to deploy a managed resource. These need to be annotated by Spinnaker as well. * fix(aws/tags): ensure tag sync is always performed during allow launch * fix(search): allow SearchProvider filter exclusion (spinnaker#1883) * allows search providers to specify an exclusion list for filter keys to optionally be removed before executing a search. * feat(provider/aws): Remove ability to automatically remove dependencies when deleting a security group (spinnaker#1894) * fix(provider/aws): trim whitespace when newlines are detected (spinnaker#1844) * fix(google): Add statusCode tag to google API metrics (spinnaker#1886) This also refactors all the tests involving google API metrics. * fix(provider/kubernetes): Add missing credential props (spinnaker#1897) * feat(provider/kubernetes): add daemon & stateful set support * fix(provider/aws): ensure STSAssumeRoleSessionCredentialsProvider get the right endpoint (spinnaker#1888) * feat(provider/kubernetes): v2 generic on demand caching (spinnaker#1884) * refactor(provider/kubernetes): v2 include api version in resource id (spinnaker#1901) * feat(provider/docker): add catalogFile option (spinnaker#1890) adds the `catalogFile` option to the Docker Registry provider. This is an attempt to make it easier to use registries that don't support `v2/_catalog` (which happens to be a lot). This file can be managed by an external process. Each caching cycle, Clouddriver reads this file to determine the list of repositories it should cache tags for. the format of the `configFile` is: ``` { "repositories":[ "library/nginx", "library/busybox" ] } ``` * fix(provider/docker): fix client constructor (spinnaker#1903) Bug introduced in `f70d6ea`. I updated the constructor to include `catalogFile` but didn't update the constructor for those using `password` and `passwordFile` configurations. This fixes that potential issue. * feat(provider/kubernetes): v2 Cache lb & scg relationships (spinnaker#1902) * Revert " fix(provider/aws): ensure STSAssumeRoleSessionCredentialsProvider get the right endpoint (spinnaker#1888)" (spinnaker#1904) This reverts commit 828fbb7. * feat(provider/kubernetes): v2 version deployed manifests as resources (spinnaker#1905) * refactor(provider/kubernetes): V2 swap kind & version in cache key (spinnaker#1907) For consistency with naming & other resources. * fix(provider/aws): Clone spot price if not explicitly specified. (spinnaker#1908) * refactor(provider/kubernetes): Segregate k8s by provider version (spinnaker#1909) * feat(provider/kubernetes): Add proper support for source capacity. (spinnaker#1906) Fully addresses: spinnaker/spinnaker#1657 Also accounts for the possibility that the user might want to set their own targetSize while ignoring the source capacity. Deck work is still needed to expose "useSourceCapacity" in the UI. Fix cross namespace server groups * feat(provider/kubernetes): Register k8s moniker for manifests (spinnaker#1910) * feat(provider/kubernetes): V2 cluster provider (spinnaker#1911) * refactor(core): s/version/providerVersion (spinnaker#1914) Having the account property named 'version' was confusing the concept of the version the provider was at, with the version the cloud environment was at. * refactor(provider/kubernetes): Rename duplicate classes (spinnaker#1913) * feat(moniker): Adds moniker to all clusters & server groups (spinnaker#1915) * chore(javadocs): Fix lt usage (spinnaker#1916) * feat(provider/kubernetes): Attach k8s annotation monikers to v2 (spinnaker#1917) resources * fix(aws): CopyLastAsgAtomicOperation instance monitoring. When not explicitly specified, instance monitoring would default to true due to groovy truth casting a non null object to true instead of traversing into its enabled attribute * fix(web): Guard against not finding a ClusterProvider (spinnaker#1921) * feat(search): add fallback query param to search (spinnaker#1912) * in an effort to speed up searching, this adds logic that attempts to pick a "good" value to use as a search query string rather than just using the empty string and pulling all the keys for a given type. * feat(provider/kubernetes): V2 Cluster details (spinnaker#1918) * chore(provider/kubernetes): Use version 0.2.0 of client-java (spinnaker#1923) * fix(provider/kubernetes): Advanced targetSize use case. (spinnaker#1920) The elvis operator before was not evaluating truthiness correctly for a value of 0. A User could and may want to deploy an update ReplicaSet of 0 and Scale it later. Or perhaps it's bundled in a Spinnaker Pipeline with multiple Server Groups and one of the Server Groups is scaled to 0. If the User is implementing 'useSourceCapacity' we would want that 0 value respected and not overwritten with a value of 1. The proposal is that the desired capacity trumps targetSize. It's reasonable to assume that someone who wants a targetSize of 0 would also have a desired capacity of 0. However, we shouldn't use the targetSize if the desired capacity is greater than its value. * refactor(provider/kubernetes): v2 cache view lives in cache package (spinnaker#1926) * feat(provider/kubernetes): v2 cache network policies (spinnaker#1927) * feat(provider/kubernetes): Build v2 creds based on context & kubeconfigFile (spinnaker#1925) * feat(provider/kubernetes): Integrate moniker into v2 (spinnaker#1929) * fix(provider/kubernetes): v2 guard against empty cache relationships (spinnaker#1932) * fix(provider/kuberentes): v2 guard against non-existent clusters (spinnaker#1934) * feat(moniker): adds moniker to server group view model (spinnaker#1936) * feat(provider/kubernetes): Cache v2 resources as artifacts (spinnaker#1931) * fix(all): Broken k8s server group lookup breaking all components (spinnaker#1937) @lwander I'm merging for short term fix since everyone is affected. However see my question about returning null and whether there are other implications. * feat(provider/google): Adds support for multiple named ports in load balancers. (spinnaker#1930) * refactor(provider/kubernetes): refactor artifact gen (spinnaker#1938) * feat(moniker): Adds moniker to pendingOnDemandResults (spinnaker#1940) * fix(provider/google): Fix named port defaults in server group deploy. (spinnaker#1939) * feat(moniker): Add monikers to security groups & load balancers (spinnaker#1941) * feat(provider/google): Support connection draining for LBs (spinnaker#1943) * fix(provider/gce): s/it/cacheData/ in zonal svg caching agent. (spinnaker#1944) * feat(provider/docker): Added insecure registry support (spinnaker#1887) * feat(provider/docker): Added insecure registry support Added an insecureRegistry flag to clouddriver-docker registry support that will skip SSL certificate validation. This will allow for addressing spinnaker/halyard#680. * Update DockerRegistryNamedAccountCredentials.groovy * Update DockerRegistryClientSpec.groovy * feat(appengine): Deploy from Google Cloud Storage [accounts]. (spinnaker#1935) * fix(provider/docker): Assign proper Spinnaker/<version> user agent (spinnaker#1946) * feat(provider/kubernetes): Properly version all versioned resources (spinnaker#1945) * fix(rollbacks): support for tolerating some instance failures (spinnaker#1922) * fix(provider/gce): Paginate server groups when calculating next name. (spinnaker#1948) * fix(appengine): Dont assume storage.gce.enabled=true (spinnaker#1953) * fix(provider/kubernetes): fix k8s client configuration and image id p… (spinnaker#1933) * fix(provider/kubernetes): fix k8s client configuration and image id problem. * (feat/cats) Add support for an error interval for clustered agent schedulers (spinnaker#1895) * feat(search): cull missing ASGs before returning (spinnaker#1951) * removes ASGs that show up in the search results from the redis cache but don't actually exist. * fix(provider/kubernetes): Fixed an issue with jobProvider failing out other providers. (spinnaker#1690) * fix(provider/dcos): fix cluster provider to guard against errors from invalid ids (spinnaker#1928) * feat(core): Add provider version to creds controller (spinnaker#1958) * feat(appengine): Add AppengineStorageController to list accounts (spinnaker#1957) * chore(provider/kubernetes): Bump client to 1.0.0-beta1 (spinnaker#1961) * fix(provider/kubernetes): v1 client call to list pods (spinnaker#1963) * fix(provider/aws): Return load balancers for all target groups associated with an application (spinnaker#1955) * refactor(provider/kubernetes): V2 deploy 1 manifest at a time (spinnaker#1965) * feat(provider/gae): Modify deploy description and validators for artifacts. (spinnaker#1966) * feat(provider/kubernetes): Feature-flag debug mode (spinnaker#1968) * fix(provider/kubernetes): Don't store-unowned artifacts (spinnaker#1969) * feat(provider/kuberentes): Application provider (spinnaker#1971) * fix(provider/aws): fix issue with VPCZoneIdentifier containing deleted subnets (spinnaker#1972) * chore(provider/docker): Only log trailing whitespace warning during failure (spinnaker#1973) * feat(provider/kubernetes): V2 cluster provider finishing touches (spinnaker#1974) * fix(provider/aws): Allow updating healthy threshold count on target groups (spinnaker#1977) * feat(moniker): Add getServerGroup endpoint that does not require application name. (spinnaker#1967) * feat(provider/kubernetes): v2 instance counts (spinnaker#1975) * feat(provider/kuberentes): Register kind mapping using autowired deploy (spinnaker#1978) * refactor(provider/kubernetes): Move manifest desc. into package (spinnaker#1979) * refactor(provider/kubernetes): Move ops w/ kind into v1 only (spinnaker#1980) * feat(provider/kubernetes): Support resize statefulset operation. (spinnaker#1976) * feat(dynomite): hashtags & dynomite pipelines (spinnaker#1832) * refactor(provider/google): Rely on cached instance details instead of calling platform MIG.listInstances() api. (spinnaker#1983) * feat(provider/amazon): Create app elb specific security group for new ELBs (spinnaker#1982) - Autocreate elb security group and ingress on application security group * feat(core/search): Improve search speed: faster validation that search keys exists (spinnaker#1984) * feat(provider/kubernetes): Support resize operations (spinnaker#1981) * feat(provider/kubernetes): V2 instance state (spinnaker#1986) * fix(provider/kubernetes): Fix v2 operation converter (spinnaker#1987) * feat(provider/gce): Replace instanceTemplate get with list call. (spinnaker#1990) * refactor(provider/kubernetes): use deployed names in result (spinnaker#1989) * fix(core): Support the eviction of stale cache key identifiers (spinnaker#1985) This PR attempts to clean up situations where the `:members` set for a given cached type contains keys that no longer exist. In particular, a `CachingAgent` can optionally return a pattern glob for each authoritative type. Any `:members` key matching the pattern glob BUT not included in the most recent caching cycle will be evicted. This has been enabled on a select few `aws` caching agents. * fix(aws): Do not do stale key evictions of amazon load balancers (spinnaker#1993) Seeing some mixed results when application/network and classic load balancers are being used. * feat(provider/kubernetes): v2 on-demand manifest caching (spinnaker#1988) * feat(provider/aws): add executionId to user-agent for cloudtrail (spinnaker#1992) * fix(aws): Enable stale key evictions for amazon load balancers (spinnaker#1994) Key structures vary between classic and application/networking load balancers. * feat(provider/kubernetes): Re-encode api types (spinnaker#1991) * feat(logging) compile logstashEncoder for json logging (spinnaker#1995) * fix(provider/gae): Set path correct when application root non-empty. (spinnaker#1970) * chore(dependencies): Bump spinnaker-dependencies to 0.115.0. (spinnaker#1996) * fix hidden exception while pulling docker tags (spinnaker#1708) * fix(provider/kubernetes): reads correct fields from cache refresh (spinnaker#1998) * feat(provider/kubernetes): v2 delete operation (spinnaker#1999) * feat(provider/dcos): Adding proxy for marathon-client to make use of spectator. (spinnaker#1960) * fix(appengine): fix deploy failures (spinnaker#2001) * feat(provider/kubernetes): Supply app name in attributes (spinnaker#2003) * feat(provider/kuberentes): implement capacity (spinnaker#2004) * feat(cats): Adding cache gzip compression (spinnaker#2000) * feat(provider/kubernetes): Update deployable's patch behavior (spinnaker#2006) * feat(provider/kubernetes): Show more in the error message (spinnaker#2007) * feat(provider/kubernetes): V2 Cache services (spinnaker#2009) * fix(web): Fixing missing property exception (spinnaker#2008) * fix(cats): Get bytes in consistent chatset during decompress (spinnaker#2010) * fix(cats): Remove use of unmodifiable collections (spinnaker#2011) * feat(provider/kubernetes): cache non-obvious relationship (spinnaker#2012) * feat(provider/kubernetes): load balancer -> sg relationship (spinnaker#2013) * refactor(aws): Simplified ELB security group autocreation (spinnaker#1997) - updated code that allows ingress from elb group to app security group * fix(provider/gae): Delete gcs target directory in deploy. (spinnaker#2015) * fix(provider/gae): Delete gcs target directory in deploy. * fix(storage): Change the storage account endpoint to '/storage'. * feat(provider/kubernetes): Enable setting of labels on k8s Service objects spinnaker/spinnaker#2035 (spinnaker#2014) * fix(provider/kubernetes): allows dots in secretNames (spinnaker#2017) The kubernetes api allows dots in secret names: https://github.com/kubernetes/kubernetes/blob/e5ac41331119bb8c67004b8efaa0c3d942b31d3a/pkg/api/validation/validation.go#L160-L162 https://github.com/kubernetes/kubernetes/blob/8d7d7a5e0d4d7e75f5a860574346944b8cc0fc43/staging/src/k8s.io/apimachinery/pkg/util/validation/validation.go#L126-L142 https://github.com/kubernetes/kubernetes/blob/e5ac41331119bb8c67004b8efaa0c3d942b31d3a/pkg/api/validation/validation.go#L1488 This allows dots in the secretName field of volume source definitions. * perf(cluster): Avoid prematurely fetching full server groups (spinnaker#2019) This is an optimization for fetching target server groups by `CURRENT`, `PREVIOUS`, `OLDEST` and `FAIL`. For these strategies, the unexpanded server group is sufficient to narrow down to a match. Once a match is found, _only_ that server group will be expanded. This only works for cluster providers that support loading both expanded and unexpanded server groups (ie. aws). The behavior for other providers should be unchanged. * fix(aws): Fix NPE in elb group autocreation (spinnaker#2020) - fixed NPE due to missing credentials in security group description request * feat(provider/kubernetes): Enable annotations, labels and secrets for security groups (aka ingress resources) (spinnaker#2000) (spinnaker#2005) * feat(google): Instrument individual google API calls. (spinnaker#2016) * fix(cats): Remove unmodifiable collections from modifiable codepaths (spinnaker#2022) * fix(provider/gce): Fix broken deploys with namedPorts. (spinnaker#2023) * fix(aws): Only autocreate app elb security group on Create (spinnaker#2025) - Updated to only auto create <appname>-elb group for new ELBs * feat(provider/aws): Support specifying explicit subnet ids for deploy (spinnaker#2026) Expectation is that a `subnetType` will still be provided and that `subnetIds` will be a subnet of those valid for `subnetType`. When explicit subnet ids are provided (or inherited), the newly created server group will be tagged with: `SPINNAKER_SUBNET_ID_OVERRIDE`: `"comma-separated list of subnet ids"` * fix(provider/gce): Paginate instance template list in svg caching. (spinnaker#2027) * fix(eureka): fix instanceId on eureka caching The instanceId was intended to be the value in dataCenterInfo.metadata but due to some Jackson behaviour the top level instanceId from the discovery registration was written over the dataCenterInfo.metadata during deserialization. This changes to explicitly handle the top level instanceId value in the JsonCreator factory method to prevent the unhandled property behaviour from kicking in. * Updating API for IAM role dropdown to work nicely with Deck
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The elvis operator before was not evaluating truthiness correctly for a value
of 0. A User could and may want to deploy an update ReplicaSet of 0 and Scale
it later. Or perhaps it's bundled in a Spinnaker Pipeline with multiple Server
Groups and one of the Server Groups is scaled to 0. If the User is implementing
'useSourceCapacity' we would want that 0 value respected and not overwritten
with a value of 1.
The proposal is that the desired capacity trumps targetSize. It's
reasonable to assume that someone who wants a targetSize of 0 would also
have a desired capacity of 0. However, we shouldn't use the targetSize
if the desired capacity is greater than its value.