How to avoid microservice dependency without slowing down your release process
We are deploying services and its a microservice based project but since we are having many small services going through parallel development we are struggling for deployment.
Example -
Admin build - v1.1.1
this is having dependancy on some other build v.x.x.x
so once build x is not fully tested we can not release admin v1.1.1
is there any solution for it to make deployment process more smooth ?
2 Pointers I can give you are:
Always have your API's backwards compatible - this way you lower your dependencies in one version or the other, creating higher isolation.
Use feature flags - when you change behaviour you can then deploy the code but it doesn't have to be active at that exact time. Use feature flags to wrap your changes and switch them on when you are ready.
Related
I am currently trying to figure out how to handle version control with microservices.
From what I have read the best strategy is to have a separate git repository for each microservice.
However when it comes to deployment having to upload multiple git repositories seems pretty complex.
Specifically I am scratching my head as how I would deploy an update where multiple microservices require changes that depend on each other, and how to roll back to the previous versions should there be an issue with a production deployment.
This seems like a headache that most developers who use micro services have had to deal with.
Any advice would be greatly appreciated, especially if this could be done with an existing library rather than building something from scratch,
thanks,
Simon
There is no easy answer or library that could solve the problem, however there are strategies that can help. I have outlined a few below
Backward compatibility of service - Whenever you are releasing make sure that your API (REST or otherwise) works with previous consumer, this could be done by proving default values for the newer attributes.
Versioning of API - When changes you are making are not small and breaking, introduce the new version of API so that older consumers can continue to work with previous version.
Canary Deployment - When you deploy a new version of micro-service route only a small percentage of calls to the new service and rest of previous version.Observe the behavior and rollback if required.
Blue Green deployment - Have two production environment, one blue which is proven working and other green which is staging containing the latest release. When the testing is done green environment and you have enough confidence, route all the calls to green.
References
Micro-services versioning
Canary deployment
Blue green deployment
Here's a plugin I wrote using some references: https://github.com/simrankadept/serverless-ssm-version-tracker
NPM package:
https://www.npmjs.com/package/serverless-ssm-version-tracker
The version format supported is YYYY.MM.DD.REVISION
In my current team we are running close to 70-80 micro services on cloud. The challenge what we face during each release is managing the dependency build versions of each service. Not all micro service
have changes for each release so we cannot update the versions for each service blindly.
It takes considerable amount of time to understand the current scope of release and change the dependency versions for only those services leaving rest unchanged. Is there a better way of managing the dependencies for releases. I am looking mainly for any open source tool or process which can be used for our team.
We are building a SaaS application. I don't have (for now - for this app) high demands on availability. It's mostly going to be used in a specific time zone and for business purposes only, so scheduled restarting at 3 in the morning shouldn't be a problem at all.
It is an ASP.NET application running in mono with the fastcgi server. Each customer will have - due to security reasons - his own application deployed. This is going to be done using docker containers, with an Nginx server in the front, to distribute the requests based on URL. The possible ways how to deploy it are for me:
Create a docker image with the fcgi server only and run the code from a mount point
Create a docker image with the fcgi server and the code
pros for 1. would seem
It's easier to update the code, since the docker containers can keep running
Configuration can be bundled with the code
I could easily (if I ever wanted to) add minor changes for specific clients
pros for 2. would seem
everything is in an image, no need to mess around with additional files, just pull it and run it
cons for 1.
a lot of folders for a lot of customers additionally to the running containers
cons for 2.
Configuration can't be in the image (or can it? - should i create specific images per customer with their configuration?) => still additional files for each customer
Updating a container is harder since I need to restart it - but not a big deal, as stated in the beginning
For now - the first year - the number of customers will be low and when the demand is low, any solution is good enough. I'm looking rather at - what is going to work with >100 customers.
Also for future I want to set up CI for this project, so we wouldn't need to update all customers instances manually. Docker images can have automated builds but not sure that will be enough.
My concerns are basically - which solution is less messier, maybe easier to automate?
I couldn't find any best practices with docker which cover a similar scenario.
It's likely that your application's dependencies are going to be dependent on the code, so you'll still have to sometimes rebuild the images and restart the containers (whenever you add a new dependency).
This means you would have two upgrade workflows:
One where you update just the code (when there are no dependency changes)
One where you update the images too, and restart the containers (when there are dependency changes)
This is most likely undesirable, because it's difficult to automate.
So, I would recommend bundling the code on the image.
You should definitely make sure that your application's configuration can be stored somewhere else, though (e.g. on a volume, or accessed through through environment variables).
Ultimately, Docker is a platform to package, deploy and run applications, so packaging the application (i.e. bundling the code on the image) seems to be the better way to use it.
After having just started using Microsoft Release Management, I am more and more convinced that it is not well suited to run integration tests. This might be a false feeling I'm having, and I'd love to get more input on this. When we first considered it, I had the intention to run the tests defined in our test plan through it's pipeline, but now I'm seeing that we should be running those as frequently as possible. We would like to run integration testing every night, but our release candidates are only defined at the end of sprints, so using Release Management for that seems conflicting.
With the tool out of the equation, we are considering exploring the Lab Template again. We did some very minor tests with it a few months ago in a legacy project but never went too far. My main concern now is that both stages need deployment:
the Release Management pipeline needs to deploy our projects to the QA and production environment
the Lab Template also needs to deploy the project on a few virtual machines to run integration tests on
The Release Management uses some very nice abstractions to achieve that. You can code machine scopes and define components based on the drop folder structure to define each part of the whole application to be deployed. On the other hand, the lab management workflow does not support this (or perhaps I'm just missing it). The standard way to make deployment work for lab testing, is to write a custom power shell script that moves the files from the build drop folder to the correct places, creates the application pools for web projects, and stuff like that, all by hand.
Ideally, I'd like to just share the entire deployment workflow between both tools and, since the Release Management way of doing it seems much simpler, I'd use that. This would make it easier to maintain both pipelines at the same time, which I assume is actually commonplace.
What is the correct approach to share the deployment code as much as possible between the two tools?
I would expect that better integration between RM and MTM/LM will be a future feature. In the interim, you could investigate using Desired State Configuration to handle having a single script that configures environments for you.
DSC support isn't really out-of-the-box in RM Update 2, but RM Update 3 will have built-in support for DSC to both Azure and on-prem VMs. Update 3 CTP 1 is out right now, but it's not production-ready.
You can still use DSC from RM in Update 2, it just requires a bit more work.
I am setting up an automated deployment environment for a number of decoupled services that are in active development. While I am comfortable with the automated deployment/configuration management aspect, I am looking for strategies on how best to structure the deployment environment to make things a bit easier for developers. Some things to take into consideration:
Developers are generally building web applications, web services, and daemons -- all of which talk to one another over HTTP, sockets, etc.
The developers may not have all running on their locally machine, but still need to be able to quickly do end to end testing by pointing their machine at the environment
My biggest concern with continuous deployment is that we have a large team and I do not want to constantly be restarting services while developers working locally against those remote servers. On the flip side, delaying deployments to this development environment makes integration testing much more difficult.
Can you recommend a strategy that you have used in this situation in the past that was worked well?
Continuous integration doesn't have to mean continuous deployment. You can compile/unit test/etc the code "continuously" thoughout the day without deploying it and only deploy at night. This is often a good idea anyway - to deploy at night or on demand - since people may be integration testing during the day and wouldn't want the codebase to change out from under them.
Consider, how much of the software can developers test locally? If a lot, they shouldn't need the environment constantly. If not a lot, it would be good to set up mocks/stubs so much more can be tested on a local server. Then the deployed environment is only needed for true integration testing and doesn't need to be update constantly throughout the day.
I'd suggest setting up a CI server (Hudson?) and use this to control all deployments to both your QA and production servers. This forces you to automate all aspects of deployment and ensures that the are no ad-hoc restarts of the system by developers.
I'd further suggest that you consider publishing your build output to a repository manager like Nexus , Artifactory or Archiva. In that way deployment scripts could retrieve any version of a previous build. The use of a repository manager would enable your QA team to certify a release prior to it's deployment onto production.
Finally, consider one of the emerging deployment automation tools. Tools like chef, puppet, ControlTier can be used to further version control the configuration of your infrastructure.
I agree with Mark's suggestion in using Hudson for build automation. We have seem successful continuous deployment projects that use Nolio ASAP (http://www.noliosoft.com) to automatically deploy the application once the build is ready. As stated, chef, puppet and the like are good for middle-ware installations and configurations, but when you need to continuously release the new application versions, a platform such as Nolio ASAP, that is application centric, is better suited.
You should have the best IT operation folks create and approve the application release processes, and then provide an interface for the developers to run these processes on approved environments.