Messagebroker feature - #807
Merged
Merged
Conversation
lars-petter-hauge
force-pushed
the
messagebroker-feature
branch
from
August 6, 2026 07:46
6efdb37 to
6a3306f
Compare
lars-petter-hauge
marked this pull request as ready for review
August 6, 2026 10:46
lars-petter-hauge
force-pushed
the
messagebroker-feature
branch
3 times, most recently
from
August 7, 2026 13:22
eec9c9c to
793891c
Compare
lars-petter-hauge
force-pushed
the
messagebroker-feature
branch
5 times, most recently
from
August 17, 2026 13:36
4c80ff3 to
9810dc6
Compare
Move the base adapter and data models into a lightweight package, add execution-free model definitions, and switch the existing API implementation to the shared contracts.
Introduce typed adapter jobs, results, heartbeats, and queue naming
Implement durable publishing, acknowledged consumption, result dead-letter topology, and explicit structural typing checks. Ensured this works with mypy for the protocols
Provide the message broker communication for each of the adapters
Make it possible to detect if there are workers alive (or if we need to wait for them to be started)
We add a listener to the results queue who will only ensure that as messages are coming in from the models, the result is persisted in the database and if necessary publish the next job.
We start all the backend services from the same location, making it obvious how many there are.
Utilise the new messagebroker setup to publish simulations from the web endpoints
Have the models run in-process instead of using web api (all but phpitz) We ensure that there is a single adapter specification, hence the workers are not depending on acidwatch_models
Add the heartbeat result in the simulations endpoint response, so that caller side can know if the simulations are being worked on or not.
lars-petter-hauge
force-pushed
the
messagebroker-feature
branch
from
August 19, 2026 07:33
6b823c4 to
c5dbb19
Compare
uv is not included in the production image, let's just run the commands from the python env directly.
lars-petter-hauge
force-pushed
the
messagebroker-feature
branch
from
August 19, 2026 07:36
c5dbb19 to
8e99f49
Compare
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
Implementing #804
Looks like a lot of changes in this one, but a lot of them are from uv files as each model has their entirely own setup. Also considering that it is a major change in architecture, it isn't as bad as it looks.
Also excluding test files:
Also, I do not intend to sneak this one in 34951dd
but we haven't required authentication for any models at all. I think we should consider this later if necessary..
A lot of these commits are preparing for change. It is really this commit that changes how the backend works, and which will then rely on a messagebroker in order to function: 7960b7b
Until this part I am actually adding packages to acidwatch backend (that is, those worker projects), and then in this commit removing all those dependencies again. I think this make the git history the most clean and easily readable (fingers crossed..)