Upgrading a target system in a continuous integration system
In an example, an apparatus may operate as a continuous delivery (CD) pipeline for application artifacts running on a target software platform. The apparatus may include a processor to operate an application upgrader for the target software platform, the application upgrader to: consume application artifacts from a continuous integration (CI) system, wherein the target software platform runs a version of the application artifacts; obtain a signed manifest output from the CI system, the signed manifest identifying the consumed application artifacts; and compare a signature of the application artifacts from the CI system to a signature of the version running on the target software platform to determine if an update of the version running on the target software platform is required. Other embodiments may be disclosed and/or claimed.
1. An apparatus comprising a memory, to operate as a continuous delivery (CD) pipeline for application artifacts running on a target software platform, the apparatus comprising:
an application upgrader executed by one or more processors for the target software platform, the application upgrader decoupling a continuous integration (CI) system from the target software platform and configured to:
periodically retrieve over a network the application artifacts that are stored in one or more artifact repositories, the CI system comprising a first pipeline to build RPMs, jars, tarballs or binary artifacts referred to as base platform artifacts, and a second pipeline to build code, configuration and other non-binary artifacts referred to as the application artifacts, wherein the target software platform runs a version of the application artifacts, and wherein the Cl system publishes the base platform artifacts and the application artifacts to the one or more artifact repositories, wherein the second pipeline is a fast content pipeline to delivery application artifacts to central cloud storage, and wherein the first pipeline is a docker driven pipeline to delivery base platform artifacts as docker images to a central container registry;
generate, for a current version of a first application artifact of the application artifacts, a signature;
prune the current version of the first application artifact by deleting hidden files to create a pruned application artifact responsive to a determination that an update is required to replace a previously deployed version of the first application artifact running on the target software platform with the current version of the first application artifact based on a comparison of the signature generated for the current version with a signature generated the previously deployed version; and
update one or more individual parts of the previously deployed version with respective parts of the pruned application artifact responsive to a determination that the one or more individual parts require updating based on a comparison of signatures of individual parts of the previously deployed version with signatures of respective parts of the pruned application artifact.
2. The apparatus of claim 1 , wherein the application upgrader is configured to run on a system for automating deployment, scaling, and management of containerized applications.
3. The apparatus of claim 1 , wherein the application upgrader is deployed as a standalone executable.
4. The apparatus of claim 1 , wherein the application artifacts comprise configuration files or binaries.
5. The apparatus of claim 1 , wherein the application upgrader is configured to: retrieve over the network from cloud storage a signed manifest output from the CI system, the signed manifest identifying the application artifacts, wherein the signed manifest is obtained from a cloud storage or a docker registry container.
6. The apparatus of claim 5 wherein the application upgrader is configured to periodically poll source endpoints to check for new application artifacts, wherein the signed manifest is obtained in response to one of the periodic polls.
7. The apparatus of claim 1 , wherein the application artifacts comprise JARS, RPMs, Images, or Compressed Zips.
8. The apparatus of claim 1 , wherein the application artifact is a machine learning model deployed on the target system for training and/or inference.
9. The apparatus of claim 1 , wherein the application updater is configured to perform the generate, compare, prune, determine, and update application artifacts up to multiple times a day.
10. The apparatus of claim 1 , wherein the application artifacts running on the target software platform provide services on the public cloud.
11. A method of providing a continuous delivery (CD) pipeline for application artifacts running on a target software platform, the method comprising:
implementing an application upgrader executed by one or more processors having access to a memory, the application upgrader decoupling a continuous integration (CI) system from the target software platform;
periodically retrieve over a network the application artifacts that are stored in one or more artifact repositories, wherein the continuous integration (CI) system comprising a first pipeline to build RPMs, jars, tarballs or binary artifacts referred to as base platform artifacts, and a second pipeline to build code, configuration and other non-binary artifacts referred to as the application artifacts, wherein the target software platform runs a version of the application artifacts, and wherein the CI system publishes the base platform artifacts and the application artifacts to the one or more artifact repositories, wherein the second pipeline is a fast content pipeline to delivery application artifacts to central cloud storage, and wherein the first pipeline is a docker driven pipeline to delivery base platform artifacts as docker images to a central container registry;
generate, for a current version of a first application artifact of the application artifact, a signature;
prune the current version of the first application artifact by deleting hidden files to create a pruned application artifact responsive to a determination that an update is required to replace a previously deployed version of the first application artifact running on the target software platform with the current version of the first application artifact based on a comparison of the signature generated for the current version with a signature generated the previously deployed version; and
update one or more individual parts of the previously deployed version with respective parts of the pruned application artifact responsive to a determination that the one or more individual parts require updating based on a comparison of signatures of individual parts of the previously deployed version with signatures of respective parts of the pruned application artifact.
12. The method of claim 11 , wherein the method is performed using a system for automating deployment, scaling, and management of containerized applications.
13. The method of claim 11 , wherein the method is performed by an application upgrader deployed as a standalone executable.
14. The method of claim 11 , wherein the application artifacts comprise configuration files or binaries.
15. The method of claim 11 further comprising: retrieving over the network from cloud storage a signed manifest output from the CI system, the signed manifest identifying the application artifacts, wherein the signed manifest is obtained from a cloud storage or a docker registry container.
16. The method of claim 15 , wherein the retrieving comprises: periodically polling source endpoints to check for new application artifacts, wherein the signed manifest is obtained in response to one of the periodic polls.
17. The method of claim 11 , wherein the application artifacts comprise JARS, RPMs, Images, or Compressed Zips.
18. The method of claim 11 , wherein the application artifact is a machine learning model deployed on the target system for training and/or inference.
19. The method of claim 11 , wherein the application artifact may be updated multiple times a day.
20. The method of claim 11 , wherein the application artifacts running on the target software platform provide service.